{ "schema_version": 1, "baseline_commit": "3519c15d7e00686a8dbf9b1e184c68726327ab1e", "baseline_source_index_sha256": "8bfdd20564f2cda54bb8264c3c802f315e3e9a11b153bd85dde4325f53c841d5", "target_new_sources": 301, "accepted_new_sources": 754, "remaining_to_target": 0, "target_count_reached": true, "current_index_counts": { "editorial_reviewed": 4, "content_screened": 2365, "abstract_checked": 1833, "metadata_only": 47, "raw_records": 4263, "unique_records": 4249, "duplicate_records": 14, "checked_sources": 4202 }, "by_category": { "Datasets": 101, "Guides and evaluations": 38, "Models": 54, "Papers": 27, "Projects": 193, "Research foundations": 341 }, "by_relation": { "adjacent": 25, "background": 341, "direct": 388 }, "by_verification_level": { "abstract_checked": 368, "content_screened": 386 }, "verification_note": "Evidence depth is labeled per source. Content screening and abstract checks do not establish full-paper review, peer review, endorsement, or experimental reproduction.", "additions": [ { "title": "AgentDecision-train", "url": "https://huggingface.co/datasets/JevAgentHarness/AgentDecision-train", "category": "Datasets", "topic": "training-dataset", "description": "Task-separated agent-state training with gold and teacher targets; raw trajectories and some training sources are omitted, with noncommercial terms stated.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "cc24f024ed1a26b1915c3332c105e9cb4ed31fd4e585eb5d699bcf03df3391ab", "retrieved_at": "2026-10-16T10:39:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "AgentDecisionBench", "url": "https://huggingface.co/datasets/JevAgentHarness/AgentDecisionBench", "category": "Datasets", "topic": "benchmark-source", "description": "Recorded agent states, answer keys and model outputs; L1 question texts are omitted, and closed-loop evaluation requires the separate harness and live models.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3f2102af399d6b88f9a37c6d4cb1c588db395b04c4c19aba2e0cb1d9198c6ba6", "retrieved_at": "2026-10-16T10:39:12+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "AIM Decision Model Browser Benchmark: 10-task sample", "url": "https://huggingface.co/datasets/AIMultiple/aimultiple-decision-models-browser", "category": "Datasets", "topic": "benchmark-source", "description": "Public browser-task sample from AIMultiple's decision-model evaluation, with task definitions and comparison provenance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e0ac5c3e9f5e795b04830e94686f7607a2f6363e89a80c93a9a31293b1f18868", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "AnyAudio-Judge Bench", "url": "https://huggingface.co/datasets/cucl2/AnyAudio-Judge-Bench", "category": "Datasets", "topic": "benchmark-source", "description": "Bilingual audio-instruction alignment benchmark with positive examples and difficult negative pairings.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "0bbedc0ccbd646b2d1242f9a5e9f7341ee108723a2255296a5b935229a2d5916", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "AnyAudio-Judge Corpus", "url": "https://huggingface.co/datasets/cucl2/AnyAudio-Judge-Corpus", "category": "Datasets", "topic": "training-dataset", "description": "Audio-judge training corpus with decomposed binary rubrics, JSON answers and per-item evidence.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "e39080509e08619b5572cd30f1f301a83b77033568d8c60b337643b4bde75daa", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Audio Jev RLCD v1", "url": "https://huggingface.co/datasets/sammyliu/audio-jev-rlcd-v1", "category": "Datasets", "topic": "training-dataset", "description": "Audio-window decision dataset with source annotations and categorical targets, separating gold from teacher-generated labels.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7b0b0c0de545e0999841ffda18f17e19f8666340230e9cc415f34ab11951b965", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "BazaarBench Analysis-v2 GPT-5 High-Reasoning Judgments", "url": "https://huggingface.co/datasets/BazaarBench/bazaarbench-analysis-v2-gpt5-judge", "category": "Datasets", "topic": "benchmark-source", "description": "Frozen BazaarBench semantic-judge run documenting the exact judge configuration and behavioral-evaluation artifacts.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "cf6d35715d846d75b7b901dbca697479d9d44bda83f605893b20c1c9ee83ab7f", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Berkeley Function Calling Leaderboard", "url": "https://huggingface.co/datasets/gorilla-llm/Berkeley-Function-Calling-Leaderboard", "category": "Datasets", "topic": "benchmark-source", "description": "Berkeley Function Calling Leaderboard's function-call evaluation cases, an adjacent tool-selection and execution reference.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "22234ad32a0bb921809278d01ed65e932b4e611690dfa17d1e0eb6ba74ab4e27", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "BFCL V1 Jev Routing Pilot", "url": "https://huggingface.co/datasets/pyuvraj/bfcl-v1-jev-routing", "category": "Datasets", "topic": "benchmark-source", "description": "Small BFCL-derived next-tool/no-function routing pilot, explicitly not a complete function-calling leaderboard benchmark.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d01a0869d27994f1a47dec32b41d21e79e8d4059012ada5db19217c15c5de9a8", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Bonzi 4B/8B vs Jev on WANLI-256", "url": "https://huggingface.co/datasets/yehor-oleksiuk/bonzi-vs-jev-wanli256", "category": "Datasets", "topic": "benchmark-source", "description": "WANLI comparison archive distinguishing independently measured smaller-model results from cited larger-model numbers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fedb22059061311e83c2677ec1e23bb7ab0d9d496ca385a55240dd4ad2e4eb18", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "calibrated-typed-decisions (v2.1)", "url": "https://huggingface.co/datasets/m-lagnajit/calibrated-typed-decisions", "category": "Datasets", "topic": "training-dataset", "description": "Typed questions retaining upstream human labels and teacher/rater-derived soft targets, including a documented correction that swaps teacher probabilities on mismatched answers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2134436aa37fc48fda8b2cdce688f7bd328f5fcf548368e197d9fd1f726bda40", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CITY-Jev: Evaluating Agent Execution Decisions Under Perturbations", "url": "https://huggingface.co/datasets/KikiNLP/CanITrustYou-Jev", "category": "Datasets", "topic": "benchmark-source", "description": "CITY-Jev perturbation benchmark for agent action selection, step judging, outcome verification and safety decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4b02d6c95169235d6b949faa81057b4827b5c14405f86798ee03ee80d4433324", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CJ-Bench", "url": "https://huggingface.co/datasets/bbldCVer-hf/CJ-Bench", "category": "Datasets", "topic": "benchmark-source", "description": "Chinese typed-decision test cases and scoring eligibility; related rows can share parent material, while contamination and cross-domain overlap remain incompletely audited.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "801ad1008ac7324fca5551fd0585b5e636ba77e9c32721068052a4fe22c8e123", "retrieved_at": "2026-10-16T10:39:12+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CJ-Dataset", "url": "https://huggingface.co/datasets/bbldCVer-hf/CJ-Dataset", "category": "Datasets", "topic": "training-dataset", "description": "Chinese typed-decision training/validation mixtures with provenance; source tasks are imbalanced, some labels are model-generated, and source-specific restrictions remain.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "14e45902d71c992a2f20e8cdf2050cb4511d310bfb20ce7d84f586f0d05d5a5f", "retrieved_at": "2026-10-16T10:39:12+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ClassOne System 1 Decision Curriculum (23,503 Examples)", "url": "https://huggingface.co/datasets/devops-thiago/classone-system-one-curriculum", "category": "Datasets", "topic": "training-dataset", "description": "Multi-domain ClassOne curriculum of structured Choice, Noul and Score questions for decision-model training.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8b2b7c1f1cc012f62d8174aeb7bed2434a4f5930f7b1ff934dde173f872e25bf", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "code-jev", "url": "https://huggingface.co/datasets/skjortan/code-jev", "category": "Datasets", "topic": "training-dataset", "description": "Typed questions about individual source-code functions for training and evaluating code-understanding decision heads.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ebffec9ea384cfdcb72eb35d7ea901c8f5702adddd3f7190aeff2210e50e3421", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Coding-agent routing decisions (Terminal-Bench 2.0)", "url": "https://huggingface.co/datasets/JevAgentHarness/coding-agent-decisions", "category": "Datasets", "topic": "benchmark-source", "description": "Coding-agent routing states and outcomes on Terminal-Bench; this is a single-trial-per-task/arm run archive, not a newly independent task benchmark.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "46ebfc7c0b87ecb7a0ebb103b57b5dd18206bc2ce47de83c3efbc273a9876add", "retrieved_at": "2026-10-16T10:39:12+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Confidence statistics for Jev and a self-judging Gemma 4 E2B", "url": "https://huggingface.co/datasets/clduab11/jev-calibration-statistics", "category": "Datasets", "topic": "benchmark-source", "description": "Aggregate Jev and self-judging Gemma confidence statistics from a retrieval benchmark, including unsuccessful pipeline results.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "213890c9fdc11dd5a61a802aa94db0e6bd743f4b740af933165c33755d3a8fdc", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Darwin-27B-JEV — Decision Index run", "url": "https://huggingface.co/datasets/FINAL-Bench/Darwin-27B-JEV-decision-index", "category": "Datasets", "topic": "benchmark-source", "description": "Archived Darwin decision-model run on a pinned Decision Index suite, including requests, results and reproduction metadata.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e63bafdbe9104f15824fb5d3563555caa314cc75c5924c89fcd8b39477568dff", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Dataset Card for JudgeLM", "url": "https://huggingface.co/datasets/BAAI/JudgeLM-100K", "category": "Datasets", "topic": "training-dataset", "description": "JudgeLM training and validation examples for teaching models to judge paired open-ended answers.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "7f84b2aba615570b7e64bc7b05f9e25349903944d820fcccd64035475926481b", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Decida synthetic typed decisions", "url": "https://huggingface.co/datasets/helmo/synthetic-typed-decisions", "category": "Datasets", "topic": "training-dataset", "description": "Synthetic Decida training questions with probability-distribution labels across Choice, Noul and Score tasks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9c564e3357ccde2d58f13f5cb7f130cfbb1f27ab4319bab9acb2c8018796a283", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Decima System One Tasks", "url": "https://huggingface.co/datasets/amyrmahdy/decima-system-one-tasks", "category": "Datasets", "topic": "training-dataset", "description": "Multilingual synthetic typed tasks with soft labels and repeated states per task in the Jev-compatible schema.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ffcec45b5f3f64e8bafac3bc551b66a238ad08f2faeaf141cfa02abb15f6659b", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Decision Model Benchmark Usage", "url": "https://huggingface.co/datasets/SaylorTwift/decision-model-benchmark-usage", "category": "Datasets", "topic": "benchmark-source", "description": "Catalog of benchmarks used by decision-model publishers, relevant to evaluating coverage and comparison choices.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "95700e0abf3841e2d41b29bbcead53a637effe4eeba5341ce4c6daabbd20e02b", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Decision models on Core AI, 231 public items", "url": "https://huggingface.co/datasets/mlboydaisuke/coreai-decision-models-public231", "category": "Datasets", "topic": "benchmark-source", "description": "Recorded Core AI model runs on a fixed public JevBench subset, with local execution and harness provenance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3fc4eb042a83b994d9ed79a90a37edde8858fd9ea76df8189906730f1754ba69", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Decision-Model Benchmark 2026-09-27: jev vs laya vs Julia-1-ONNX", "url": "https://huggingface.co/datasets/morrenhale/decision-benchmark-jev-laya-julia", "category": "Datasets", "topic": "benchmark-source", "description": "Frozen manifests and raw outputs comparing Jev, Laya and Julia on identical decision questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "157072741cdf6b3349ed451de4f674e0419945d154f6030b86b86783f937d40b", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "decision-models-as-guardrails", "url": "https://huggingface.co/datasets/raxITLabs/decision-models-as-guardrails", "category": "Datasets", "topic": "benchmark-source", "description": "Guardrail evaluation cases for harmful content, prompt attacks, denied topics, filtering, PII and grounding decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "53d54a9d73c712ea513f7f1be59d5e0aa27474fcf7292306b16dd30486585a7d", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "DeRef tool-call argument verification sets", "url": "https://huggingface.co/datasets/lm2445/DeRef", "category": "Datasets", "topic": "tool-argument-evaluation-data", "description": "Argument-level evaluation sets with correctness/path labels and a source-relocation control; derived from existing benchmarks and model trajectories, not an independent new task population.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9205ef0d7e00e754314694857f1227ee4c7206141df0fd19a6772929b1eed29c", "retrieved_at": "2026-10-16T11:23:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "EdgeIntent v1", "url": "https://huggingface.co/datasets/UTSCybeR/Edge-Computing-JEV", "category": "Datasets", "topic": "benchmark-source", "description": "EdgeIntent benchmark pairing edge-service requests with typed intent contracts for orchestration studies.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "027ae03c8104549a481317cde08a128f2c5e6354e6e40953d15bd92d91a9c20d", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "fomc original source annotations", "url": "https://huggingface.co/datasets/social-jev/fomc", "category": "Datasets", "topic": "training-dataset", "description": "Original FOMC annotation records for social-science decision studies, with missing labels explicitly retained as unlabeled.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a79ad4018d051892028caa3ffc1892848b3ed3e272f011684aeb44c78027764e", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev 1.13 on OpenTWBench (self-reported)", "url": "https://huggingface.co/datasets/kangjwme/jev-opentwbench-results", "category": "Datasets", "topic": "multilingual-evaluation-results", "description": "Traditional-Chinese benchmark result archive recording option permutations and mapped predictions; question text remains in the upstream labeled datasets.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "339e6c0719201c36e44636ec12f3cb2638491ec3edba4f260504c2a4c44ad438", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev Decisions v1", "url": "https://huggingface.co/datasets/samatv256/jev-decisions-v1", "category": "Datasets", "topic": "training-dataset", "description": "Decision-oriented corpus derived from public agent trajectories for tool selection, routing and local agent control.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "70c90a2ae32f100d91d2ee28d89f983b2463035dbdeca31809990dabbfaeb0bb", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev evaluation traces", "url": "https://huggingface.co/datasets/Ayushnangia/jev-evaluation-traces-20261008", "category": "Datasets", "topic": "benchmark-source", "description": "Retired ToolEmu and AgentDojo raw evaluation traces, retaining failures and partial runs rather than claiming validated final scores.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1f2253496176e0f568fb44253d2d72d4af5f6408a2975d49968a4777ac924c2a", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev Logs log-triage benchmark", "url": "https://huggingface.co/datasets/reachjalil/jevlogs-log-triage-benchmark", "category": "Datasets", "topic": "benchmark-source", "description": "Sanitized public-log triage benchmark evaluating whether log lines should be escalated to a reasoning model.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1590bb68c807c67b47d1feaca9ca0b0f40c15769b4d737d431d4e19543425a58", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev RAG Benchmark (English)", "url": "https://huggingface.co/datasets/emretheus/jev-rag-benchmark", "category": "Datasets", "topic": "benchmark-source", "description": "Frozen-candidate RAG comparisons of Jev reranking, a cross-encoder and no reranker, with per-query artifacts.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "937d947e512adb4e917c26ec0fc632cc261497bcbd9233816cdc142f2deb512e", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev RAG Chinese Benchmark (S1)", "url": "https://huggingface.co/datasets/cnfeichen/jev-rag-zh-bench", "category": "Datasets", "topic": "multilingual-evaluation-results", "description": "Chinese reranking run archive on public hard-negative retrieval queries, comparing Jev question primitives with a lexical baseline on shared relevance labels.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ac64b9376e970c941768720d81518b24fc537c65e96dbc2d504ba382f840407f", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JEV robotics pre-training dataset", "url": "https://huggingface.co/datasets/yuyanghu06/jev-27b-vla-pretraining", "category": "Datasets", "topic": "training-dataset", "description": "Synthetic robotics-control replay data for a Jev-like project, with source provenance and stated physical-validation limits.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9b093389865fb89875fda499da8708f087fb50048719fd3aee5fe242aa99f904", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev vs Luna as a PagerDuty trigger", "url": "https://huggingface.co/datasets/reachjalil/jev-luna-pagerduty-trigger", "category": "Datasets", "topic": "benchmark-source", "description": "Synthetic alert stream comparing Jev and a reasoning model as PagerDuty triggers, without production traffic claims.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a7698aa9cd60f20d60d5fcc1a19b29212a1a5e5d816d031638ac7d224891f3f7", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev-AR Bench", "url": "https://huggingface.co/datasets/atmaneayoub/jev-ar-bench", "category": "Datasets", "topic": "benchmark-source", "description": "Arabic intent-routing evaluation across standard, dialectal and mixed language slices, plus unseen-domain RAG tests.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c085eebd34c8f9092fd9afd53bd4785540d09af8093ef3b242ff80fdbf0fed5e", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jev-bench", "url": "https://huggingface.co/datasets/hayriyigit/jev-bench-tr", "category": "Datasets", "topic": "benchmark-source", "description": "Turkish machine translation of Jev-bench with preserved labels, pinned source revision and translation manifest.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "706fb1d03d524dc990197f3cd2a4bd29ac24b4bc6978059a908c6bcca36f3c7d", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jev-bench", "url": "https://huggingface.co/datasets/kishida/jev-bench", "category": "Datasets", "topic": "benchmark-source", "description": "Multiple-choice evaluation set designed to score option probabilities rather than generated answer strings.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2864ab2771df23d2fc61f7d8101dde8010dfd781b7500067b4da70fad4ee6044", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev-class: name the weakness class", "url": "https://huggingface.co/datasets/trapstreet/jev-class-cve-bench", "category": "Datasets", "topic": "benchmark-source", "description": "CVE-to-CWE finite-choice benchmark studying vulnerability-class selection with a larger candidate set.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f428808082e354ee380f167aab10c00720d60e27b4d82446087ccfe9daab9944", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JEV-Contrast", "url": "https://huggingface.co/datasets/tudo-vn/JEV-Contrast", "category": "Datasets", "topic": "multimodal-robustness-benchmark", "description": "Visual-choice benchmark with source-derived labels and model-written distractors, not fully human-verified; both option lists place the gold answer first and require shuffling with label remapping.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7384f4309e713b1e91bfc3b103130321f57610c1721296f064378a5bc88a6eb0", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jev-my-bro Governance Dataset", "url": "https://huggingface.co/datasets/JonusNattapong/jev-my-bro-dataset", "category": "Datasets", "topic": "training-dataset", "description": "English-Thai governance decision cases about execute, ask-user, reject, review and risk judgments.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "38f2f747cf8f8b0e497036245efe21b4ab758748da374a987e5078dafd7c8e03", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev-style Structured Decision Dataset", "url": "https://huggingface.co/datasets/GhostScientist/jev-decisions-v1", "category": "Datasets", "topic": "training-dataset", "description": "Synthetic structured states and labeled options for training single-letter Tev/Jev-style decision models.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9d40d013798c39e6da91106a2ab60f567507e0c41c7999a93c358193e706b2d0", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev-tree vs the 255 choice cap", "url": "https://huggingface.co/datasets/reachjalil/jev-tree-choice-cap", "category": "Datasets", "topic": "benchmark-source", "description": "Synthetic incident hierarchy testing tree-based selection beyond Jev's per-question choice limit.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "28935a4af78f522758c5b696aaff63ea6fa4c27f7869fea18bad785bf7d05ffd", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevAdvBench", "url": "https://huggingface.co/datasets/Hangtao/JevAdvBench", "category": "Datasets", "topic": "benchmark-source", "description": "JevAdvBench adversarial test data for typed decision models, covering clean requests and attacks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ce31c30a048c29a7272d2b9e668f3569850120a471f737704f6131a9bdcd4643", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JEVal", "url": "https://huggingface.co/datasets/carlosxiang/JEVal", "category": "Datasets", "topic": "benchmark-source", "description": "JEVal benchmark data for comparing general decision models beyond Jev, accompanying its empirical study.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "818ff20edee14db5836ab39c856ebe828daf1ae89ed55c83019df6729715c09c", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevAlt training data: text classification in English, Turkish and German", "url": "https://huggingface.co/datasets/mertkayacs/jevalt-data", "category": "Datasets", "topic": "training-dataset", "description": "English, Turkish and German decision-training data supporting the JevAlt model family.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8b1aa72a6904845c2018ffde4fcf08cb2dd235a709bc2709cd0e917c02237901", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevBench", "url": "https://huggingface.co/datasets/JevBench/jevbench", "category": "Datasets", "topic": "benchmark-source", "description": "Metamorphic coherence benchmark checking relationships among typed probability predictions rather than accuracy alone.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "309f9f1a3d05982611926fff97f28e3e174135983a9ed891bd2795aa77e52130", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevBench", "url": "https://huggingface.co/datasets/Leanmcp/jevbench", "category": "Datasets", "topic": "benchmark-source", "description": "JevBench framework cases and model predictions on matched inputs from public datasets.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e2e0c301caef35cb4a9dbe09dd89c18b6ede1c2b7101bb04d03d67764f193ed7", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevBench decision model results", "url": "https://huggingface.co/datasets/benchmarkheaven/jevbench", "category": "Datasets", "topic": "benchmark-source", "description": "Benchmark Heaven's decision-model results and scoring-method documentation for calibration, accuracy and runtime comparisons.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f33e94835a5f7dc18f0d29e3269e147d83e8c2e1baaf325ff8707860704f70e5", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevEmbed-Data", "url": "https://huggingface.co/datasets/HIT-TMG/JevEmbed-Data", "category": "Datasets", "topic": "training-dataset", "description": "JevEmbed training and evaluation questions with request/answer JSON fields and documented source splits.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8714dd09834dfee34eb8158c5e08c0f5df80092a76ac408fdd9cbfa54c7cac0e", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jevhome decisions", "url": "https://huggingface.co/datasets/jevhome/jevhome-decisions", "category": "Datasets", "topic": "training-dataset", "description": "Jevhome training corpus of typed state/question pairs with teacher-provided gold answers and probabilities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "89cb13b5abad5d2132dcddaec843e03c03587f7542f63d6a24ebe91f5a5ca07f", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevJudge-Public v0.4: Jev Models as Multimodal Agentic Judges", "url": "https://huggingface.co/datasets/SimpleJev/JevJudge-Public", "category": "Datasets", "topic": "benchmark-source", "description": "JevJudge multimodal agentic-judge evaluation data, released with evaluation-only and upstream-license restrictions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "94266340ec028d496d22dd5de854a6e7a686ab1d1266b4899178b8c762306654", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevJudge-Robotics-Public v0.1: Judging Robot Manipulation", "url": "https://huggingface.co/datasets/SimpleJev/JevJudge-Robotics-Public", "category": "Datasets", "topic": "multimodal-robustness-benchmark", "description": "Robotics-judge evaluation with blank-media, shuffled-media and permuted-option controls, disclosing human and construction-based labels plus text-shortcut risks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ca2bf9f1550d2a3c43296b07568a8ae17c6806e7005f03d67845b24c6c4f4bb2", "retrieved_at": "2026-10-16T10:38:00+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JudgeBench: A Benchmark for Evaluating LLM-Based Judges", "url": "https://huggingface.co/datasets/ScalerLab/JudgeBench", "category": "Datasets", "topic": "benchmark-source", "description": "JudgeBench test data evaluating LLM judges on objective correctness judgments.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "4574f6ad5bb5c8c2a8bd7d357413b010d860f52dcceaea8e78a5c9f4df7a929a", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Korean JEV Benchmark", "url": "https://huggingface.co/datasets/DaehanKim/korean-jev-benchmark", "category": "Datasets", "topic": "benchmark-source", "description": "Korean finite-choice benchmark assembled from public datasets, with normalized options and baseline provenance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "198ab85f3c0c76547e6f7165a0ff47b590163130ef785c58023e4b13ef647d7b", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Korean Jev Decision Index", "url": "https://huggingface.co/datasets/bzantium/korean-jev-decision-index", "category": "Datasets", "topic": "benchmark-source", "description": "Korean decision benchmark combining reasoning, language, tool use and preference tasks with probability outputs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9ed43b3a7fdd202775dc4a2bfb6e8f1473d7ac46e30e0ec4a24d1780ba8a455c", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Laya vs Jev", "url": "https://huggingface.co/datasets/Luni/laya-jev-benchmark", "category": "Datasets", "topic": "benchmark-source", "description": "Matched Laya/Jev comparison artifacts that examine limitations of cross-benchmark performance claims.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1b0fc84037c927ac361941e4907e95f57c95bd1aacc0a468ed89da2aaf511ea0", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Najd System One — public research draft", "url": "https://huggingface.co/datasets/najdresearch/system-one", "category": "Datasets", "topic": "training-dataset", "description": "English and Arabic System One research draft that explicitly notes pending independent linguistic and label review.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "3517c628f95047a260a8188dbb6f689195f1d84c8f28550d407a54a5bfd49b42", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Next JEV Stage 1 multi-model CoT", "url": "https://huggingface.co/datasets/JonesLin/next-jev-stage1-multi-model-cot", "category": "Datasets", "topic": "training-dataset", "description": "Multimodel reasoning-response export for next-JEV training; explicitly a CoT-only release without classification targets.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2f895eef418f462f2eace61a38168dc60b32dc9817ceaa401e960f3e2d734a8a", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "NOTAM typed decisions", "url": "https://huggingface.co/datasets/AirsideLabs/notam-typed-decisions", "category": "Datasets", "topic": "benchmark-source", "description": "Frozen NOTAM typed-question evaluation suites with calibration analysis; explicitly restricted from operational aviation use.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "56a34379010f4cf160676b810a563a8c2dfac99a1a8f74c1bf93fbf010a7a443", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1", "url": "https://huggingface.co/datasets/nvidia/Nemotron-RL-Agentic-Function-Calling-Pivot-v1", "category": "Datasets", "topic": "training-dataset", "description": "NVIDIA function-calling RL data reformulating expert tool-use steps as behavior-cloning decisions.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "d27b7c2df3576c97880399d979e00740e5ad698fa6e0e188371471087d90b484", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "open-system-one-bench — per-item predictions, 10,000 decisions, 5 stacks", "url": "https://huggingface.co/datasets/dylantom2012/open-system-one-bench", "category": "Datasets", "topic": "benchmark-source", "description": "Per-item classification and routing predictions comparing Jev, Laya and other stacks on matched inputs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "aaa4285dc10eddf9f01124523f725d1def4ab2b386016789fd967be21da600a1", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "openjev-rlcd-v0", "url": "https://huggingface.co/datasets/soyrsoyr/jev-playground-rlcd-v0", "category": "Datasets", "topic": "training-dataset", "description": "Procedurally generated typed questions with reference probability distributions defined by the generator.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "516528af8cd37d01932fd050d8d355427a5cf4f7187acff8cf57e9e0ba044ffd", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Pythonic Function Calling Dataset", "url": "https://huggingface.co/datasets/driaforall/pythonic-function-calling", "category": "Datasets", "topic": "training-dataset", "description": "Synthetic Pythonic function-calling examples for tool-use training, an adjacent generative baseline resource.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "25d403630b33b026c74e803b7be09906b1f57cd343af17c15d074b801d013d92", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "RANIntent v1", "url": "https://huggingface.co/datasets/UTSCybeR/6G-JEV", "category": "Datasets", "topic": "benchmark-source", "description": "RANIntent benchmark mapping natural-language network intents and telemetry to typed O-RAN policy decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3849ef6e3e393a909391a5cf1453ba0c1e2877cadcde12de6ef4f8426ba05f7a", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Results for Holisitic Evaluation of Reward Models (HERM) Benchmark", "url": "https://huggingface.co/datasets/allenai/reward-bench-results", "category": "Datasets", "topic": "benchmark-source", "description": "Published raw reward-model benchmark scores supporting reanalysis of evaluator behavior.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "15b3ad0f0b498cae879a6524be4b2309c9177cf17526de97976644274f3cf012", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Reward Bench Evaluation Dataset Card", "url": "https://huggingface.co/datasets/allenai/reward-bench", "category": "Datasets", "topic": "benchmark-source", "description": "RewardBench preference-pair evaluation data covering chat, safety and reasoning comparisons for reward models.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "819086ae803ffe3d3126eb2c188b8c67e2a0acb91b67096727f59f5aff423256", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "RewardBench 2 Evaluation Dataset Card", "url": "https://huggingface.co/datasets/allenai/reward-bench-2", "category": "Datasets", "topic": "benchmark-source", "description": "RewardBench 2's distinct harder evaluation set, including unseen human data and additional reward-model test categories.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "4f9f92600767160c6b7b1a3c822361dabfea9ac254f8a152a02cb96ab0650911", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "RLCD Decision Dataset (v1)", "url": "https://huggingface.co/datasets/anthonym21/rlcd-decision-v1", "category": "Datasets", "topic": "training-dataset", "description": "RLCD decision training data with finite-option questions and calibration-oriented supervision.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "067e4b493ecfb195b3deea059c432958aad095b20cab54568b3abfd71d4d445b", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "RSI-Jev benchmark runs", "url": "https://huggingface.co/datasets/shgao/rsi-jev-benchmarks", "category": "Datasets", "topic": "benchmark-source", "description": "Published RSI-Jev benchmark runs with raw Decision Index outputs and run documentation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2f5d2197f00717befb4951e523b9f0717a2ee937e59e975869d9067b9ce37183", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "RU-LLM-Judge-Dataset", "url": "https://huggingface.co/datasets/AtesiT/ru-llm-judge-dataset", "category": "Datasets", "topic": "training-dataset", "description": "Russian LLM-judge dataset documenting collected model judgments, relevant to multilingual evaluation supervision.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "debcb178982c4727a38c196c93e74a856b85246ea636ed738a0389832ad0b8a1", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SearchJev Dataset", "url": "https://huggingface.co/datasets/SearchJev/search_jev_unified_v2_release_20260928", "category": "Datasets", "topic": "training-dataset", "description": "SearchJev training and evaluation release for search-agent decisions, with versioned split and quality-sidecar documentation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f121531a137cf06e766db95c698ee3d2c00766c46c20b012c0ab716b7aab6263", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SII-Decision-27B Training Data", "url": "https://huggingface.co/datasets/jylin001206/SII-Decision-27B-Training-Data", "category": "Datasets", "topic": "training-dataset", "description": "Teacher-probability targets for SII-Decision training, with validation and source manifests; targets are model-derived, and upstream data licenses vary.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "63c3d0934a7dd788c6f5ed9a141874ff8918c56bac920ac565cd27e8dda6ce47", "retrieved_at": "2026-10-16T10:39:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SN92 specialists: train + dev data for 20 business use cases", "url": "https://huggingface.co/datasets/decisionmodels/sn92-specialists", "category": "Datasets", "topic": "training-dataset", "description": "Synthetic specialist-training and development decisions for multiple business workflows in a Jev-compatible format.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c5d8013a9b1def7fc69a995555ea80f52932eda98322db79a6b90ac345051cbc", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Social JEV Codebooks", "url": "https://huggingface.co/datasets/social-jev/codebooks", "category": "Datasets", "topic": "training-dataset", "description": "Social-science classification codebooks defining labels and meanings for probabilistic annotation studies.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b98cb98def16645e182b20b8a724e66a2a9a0ded2416546db6db5a151a9905d3", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SoK-JEV", "url": "https://huggingface.co/datasets/UTSCybeR/SoK-JEV", "category": "Datasets", "topic": "benchmark-source", "description": "Frozen data and results accompanying a survey of semantic decision engines in network-control loops.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f120c6c96f88d8c20d006ea8b359c3be72ad8da57bbf541d7b147985aaa89f0e", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Support Routing v1", "url": "https://huggingface.co/datasets/decisionmodelhub/support-routing", "category": "Datasets", "topic": "benchmark-source", "description": "Fictional support-ticket routing cases under a written policy, with held-out model predictions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d474614a6887c59dfd3718daf8ecff78be079fd74db9457a6b848fe0ce83c928", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Support Urgency v1", "url": "https://huggingface.co/datasets/decisionmodelhub/support-urgency", "category": "Datasets", "topic": "benchmark-source", "description": "Fictional support-urgency cases with explicit yes/no criteria and recorded model probabilities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c33093662e25104d97b1ee86e32ab6eb8772b72c901c80fedc84a9120c7b0f58", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "System One Eval", "url": "https://huggingface.co/datasets/blazeofchi/system-one-eval", "category": "Datasets", "topic": "benchmark-source", "description": "Small synthetic structured-state reasoning diagnostic; explicitly not a population benchmark or blinded leaderboard.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b990164eab65f384ee678f1063b213ca65a10795d9c021868cf26f683fd15dcf", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "System One Mini Synthetic Diagnostics", "url": "https://huggingface.co/datasets/DavidHatley/system-one-mini-data", "category": "Datasets", "topic": "training-dataset", "description": "Deterministic synthetic intervention diagnostics for System One Mini, without real agent traces or teacher-model labels.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6bff668aed0a6b5fb815025f9f2dd57950444ce41e3939a404c6c3cb85fad741", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "systemone-lite-general", "url": "https://huggingface.co/datasets/dwidlee/systemone-lite-general", "category": "Datasets", "topic": "training-dataset", "description": "Rule-labeled synthetic System One Lite data, including IID and harder layout/paraphrase test splits.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "11c19d640fba5ea2bff579a72b25e058d70a1ef4315af47807d3bb5661032219", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Tetris placement decisions with known answers (10k)", "url": "https://huggingface.co/datasets/JonesLin/next-jev-tetris-decisions-10k", "category": "Datasets", "topic": "training-dataset", "description": "Tetris placement questions with known optimal options for verifying teacher-generated reasoning.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2ed1d4f21e24b58c33100c43a39b069dbbf16698bdaacb44db9747ada68df41d", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "turkish_jev_noul", "url": "https://huggingface.co/datasets/ahmetege/turkish_jev_noul", "category": "Datasets", "topic": "training-dataset", "description": "Turkish yes/no decision questions derived from documented public language datasets in the Noul format.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "104661e8d7a27d7756a8cdde2bff124a200829d09cd19b65d0691128588504ac", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Two-Box Judge GUI Dataset", "url": "https://huggingface.co/datasets/THU-BoZhang/two-box-judge-gui", "category": "Datasets", "topic": "training-dataset", "description": "GUI screenshot tasks choosing between two candidate bounding boxes, suitable for studying bounded visual judgments.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "962d7047eaefbd0ea93a9240bd729dae0c042bba27d41a2eb0a8289fcd3bab99", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Type-Safe Models: Safety Classification and Choice Robustness", "url": "https://huggingface.co/datasets/SabaPivot/jev-safety-bench", "category": "Datasets", "topic": "benchmark-source", "description": "Safety classification and option-order robustness tests, explicitly distinguished from evaluating actual refusal behavior.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "92aa54b2737e1ed0f61a68b6596229308309a5b22b7388a3cf72d69749fcbf42", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed Decision Bench v0.3", "url": "https://huggingface.co/datasets/SamuelChien821/typed-decision-bench", "category": "Datasets", "topic": "benchmark-source", "description": "Typed Decision Bench cases and scoring documentation spanning several workflow-oriented evaluation suites.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "113eefa90fd44d61068668ecaf2d635d9ca9e84bbbca1450a513066bed87642f", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed Decisions (Deutsch)", "url": "https://huggingface.co/datasets/johannhartmann/typed-decisions-de", "category": "Datasets", "topic": "training-dataset", "description": "German translation of the typed-decisions workflow benchmark with preserved probabilistic answer structures.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7cdf57646fe795c9b4b22b5c41510dc2951b3e19ef821f5b3470ee79941599ed", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed Decisions (Russian)", "url": "https://huggingface.co/datasets/yyhlm/typed-decisions-ru", "category": "Datasets", "topic": "training-dataset", "description": "Russian translation of typed-decisions retaining original splits and probability labels for multilingual evaluation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "99160aa82e9467bfe7d3788b2a7363208a85acc051c4a69522a7f476251b7b97", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed Decisions PT/ES", "url": "https://huggingface.co/datasets/telepatia-ai/typed-decisions-pt-es", "category": "Datasets", "topic": "training-dataset", "description": "Portuguese and Spanish typed-decisions translations preserving splits, identifiers, labels and target distributions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8a1a6c0702be5eb381f979a030adba6451f8255a503c29370d7e8ae4c966f80c", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed Decisions Synth", "url": "https://huggingface.co/datasets/n4ze3m/typed-decisions-synth", "category": "Datasets", "topic": "training-dataset", "description": "Synthetic Hmm training corpus spanning many workflows, with finite-option answers and soft probability labels.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b0b1f61d4e03851bda9ce6f960639e8955affaf26b796e56eb57d2857a7ef3e6", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed Decisions — 日本語版（typed-decisions-ja v3）", "url": "https://huggingface.co/datasets/GeneLab/typed-decisions-ja", "category": "Datasets", "topic": "training-dataset", "description": "Japanese translation of the typed-decisions workflow benchmark with source revision and translation provenance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "42e05aa040a76fd26f8df4e75606a78f35554d4846c0a1346517d37780f94839", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed-Decisions benchmark results", "url": "https://huggingface.co/datasets/ML-Intern-lab/typed-decisions-benchmark", "category": "Datasets", "topic": "benchmark-source", "description": "Recorded decision-model accuracy and timing runs on the typed-decisions benchmark, with hardware and run context.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "061dbdeed1abe49445510b13c53e710d9162b86a7c5da70e509f12e8345c4413", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "typed-decisions: code-holes", "url": "https://huggingface.co/datasets/com-kotobalabs/typed-decisions-code-holes", "category": "Datasets", "topic": "training-dataset", "description": "Code-token substitution questions mined from public repository histories, with commit-derived choice labels.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b24485fb54aec85577322ce5c58d74c70413d42a64a55c02c2d76681b3136c07", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "UK Acts original annotations", "url": "https://huggingface.co/datasets/social-jev/uk-acts", "category": "Datasets", "topic": "training-dataset", "description": "UK legislative source annotations with documented codebook-version differences for classification-reproduction studies.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "205660ae95d481a68583ce7825c7f3ec90725766a87813114eeebb2f88033b38", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "uk_pmqs original source annotations", "url": "https://huggingface.co/datasets/social-jev/uk-pmqs", "category": "Datasets", "topic": "training-dataset", "description": "UK parliamentary-question source annotations with explicit codebook provenance and no generated labels.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1a9ccc7ebac0c546949216f622448694773b0438eeadd6e954ee6851c29ad1ed", "retrieved_at": "2026-10-16T10:05:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "vistalab-system-one-synthetic", "url": "https://huggingface.co/datasets/UlkuTuncerKucuktas/vistalab-system-one-synthetic", "category": "Datasets", "topic": "training-dataset", "description": "Synthetic Turkish typed questions and teacher probabilities used for training the VistaLab decision models.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "af61f42a1d38a349fb3887e34301bd699a72f60a42326f6d2a28f9894e5e27f1", "retrieved_at": "2026-10-16T10:05:26+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Visual Jev decisions v1", "url": "https://huggingface.co/datasets/harshnandwana/visual-jev-decisions-v1", "category": "Datasets", "topic": "training-dataset", "description": "Visual decision training and test records with provenance tiers; unreviewed teacher candidates are separated from accepted splits.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2ba8e3c4db7c3b874efcb7919b808e4880523fd869ae8c8c88f7293937b71533", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Xyntetik typed decisions: TD-000 to TD-002 record", "url": "https://huggingface.co/datasets/Joakimpalm-Zen/Xyntetik-Typed-Decisions-record", "category": "Datasets", "topic": "benchmark-source", "description": "Preregistered typed-probability study records, including an explicitly inconclusive fine-tuning comparison.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4dce2b4c775e42a48052f2436c14d6788a4f4414320150496a19de1a755211bf", "retrieved_at": "2026-10-16T10:05:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Zero-shot classifiers, retrieved examples and trained models on 113-class CFPB complaint labels", "url": "https://huggingface.co/datasets/earino/ecbs5200-jev-benchmark", "category": "Datasets", "topic": "benchmark-source", "description": "Evaluation artifacts for a 113-class complaint taxonomy, comparing zero-shot, retrieved-example and trained classifiers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "68a1968ddeba1d343ed38b115c6952767c4b6d88d61434089f936b9120002b4e", "retrieved_at": "2026-10-16T10:05:27+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Advanced usage", "url": "https://docs.typesafe.ai/sdk/python/usage/advanced", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official advanced Python client usage and operational patterns.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "30d88a3443a7d1c88c611c0ccd94955b9735e44bb59366aa9793852945d6b860", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Advanced: structure", "url": "https://docs.typesafe.ai/primitives/advanced", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official advanced usage of decision primitives and their interface constraints.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "650a61153fe3d16619e57c8f81fc03b570d663c40dd2fe94df6e57db6c906f62", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Agent skill", "url": "https://docs.typesafe.ai/agent-skill", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official agent integration instructions for using the TypeSafe SDK and typed decision interfaces.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c3f8411c9fbdbe4dd6c0800f1aa14acf63bd297415f3e9a9fabd8c8429c9b24d", "retrieved_at": "2026-10-16T10:07:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "AI primer", "url": "https://docs.typesafe.ai/introduction/machine-learning-primer", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official machine-learning primer for interpreting decision-model behavior and evidence.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "62f4354cfcdc00c5ec3e4865bc6b6a79900042c26c3afea0e67d88537e980121", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Artificial Analysis Decision Index | Artificial Analysis", "url": "https://artificialanalysis.ai/evaluations/artificial-analysis-decision-index", "category": "Guides and evaluations", "topic": "decision-model-evaluation", "description": "Independent hosted decision-model benchmark covering capability, calibration, cost and latency; the suite includes private questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "badc0c6ad67b6d06140fdfcc79109b2746077491cbd2a3ce44111456e969c83a", "retrieved_at": "2026-10-16T10:07:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Autoresearch feature discovery", "url": "https://docs.typesafe.ai/cookbooks/autoresearch_feature_discovery", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official feature-discovery loop using TypeSafe questions to turn free text into numeric features for supervised CatBoost regression.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "24e0854e5a5846c269142d2a57abfb8aa90d63c37231b03036ed08f1e9989b72", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Basic usage", "url": "https://docs.typesafe.ai/sdk/python/usage/basic", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official introductory Python client examples for decision requests.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "25a9e959ef68bb9c3302b99fb8456aa35782ff4ee41c7bb31bfe92e7c58221cb", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Classification using confidence", "url": "https://docs.typesafe.ai/cookbooks/classification_using_confidence", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official example of confidence-based classification; thresholds need validation on the intended workload.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6b5bed0126a9df1dac941760a57f82b5a336d866b03a0634ff4a845c4c146f35", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Decision Index Benchmarking | Artificial Analysis", "url": "https://artificialanalysis.ai/methodology/decision-index", "category": "Guides and evaluations", "topic": "decision-model-evaluation", "description": "Primary scoring and task-selection methodology for Artificial Analysis Decision Index, including private and filtered subsets.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f926ff9aea58e234c222929beeb61cf799ad65637bc89a98b9b86c8191de547f", "retrieved_at": "2026-10-16T10:07:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Example use cases", "url": "https://docs.typesafe.ai/concepts/use-case-map", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official map of tasks suited to decision primitives and the boundaries of those applications.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4dd749d672520170b99113c870ca8aff93e16212075b210c8797849d8ba24fbc", "retrieved_at": "2026-10-16T10:07:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Function calling", "url": "https://docs.typesafe.ai/cookbooks/function_calling", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official example of selecting functions with decision questions while execution remains in application code.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c03cd58ac34dca4e282bfda4b21df943f27c0d652d49ad6b8bac1afa47f8856c", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "How Jack & Jill Cut Candidate Screening Costs By 88% With Jev - TypeSafe AI Blog", "url": "https://typesafe.ai/blog/jack-jill-jev-case-study", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Vendor case study of Jev in a production workflow; provider-reported operational evidence.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6dc5aef6d947bd88d4200911491d4b202fb2f28b0fe4ff0a9ccb0afe2d8de147", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "How to build with TypeSafe", "url": "https://docs.typesafe.ai/concepts/how-to-build-with-system-one", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official guide to designing applications around bounded System One questions and program-controlled actions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9f5a20ac39383695d554bf812b365ed4b70622b8e5ac10484dce290469c68972", "retrieved_at": "2026-10-16T10:07:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Intent routing", "url": "https://docs.typesafe.ai/patterns/intent-routing", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official intent-routing pattern using explicit categories and application logic.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "94aeca51a9c36b044876850ec8ac2b23dccff60506334ad1795796bb5fa760a6", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JavaScript SDK", "url": "https://docs.typesafe.ai/sdk/javascript", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official JavaScript SDK usage entry point for typed decision requests.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e97bef8031ae1db7d1172c642e3d7f642323a130bde0acfcfde533bd66afff1a", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev in Medicine — author supplementary release", "url": "https://zenodo.org/records/23008156", "category": "Guides and evaluations", "topic": "evaluation-harness", "description": "Medical benchmark supplementary code and shareable outputs; archive contents were not inspected, copyrighted NEJM text is excluded, and the release is not clinical validation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a4e00b3719107b3670b0eb4172fc5f1f631a668f429f82d8b94c2f5654e3d674", "retrieved_at": "2026-10-16T11:57:25.427624+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev with coding agents", "url": "https://docs.typesafe.ai/introduction/coding-agents", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official integration guidance for coding agents using TypeSafe decision models.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "72b81af8a34379d9c6d0f632aabc6b69b98dad6a85371ffd088687e844841a80", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev: For the builders AND dreamers - TypeSafe AI Blog", "url": "https://typesafe.ai/blog/jev-1-14", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official Jev 1.14 release description; reported capability changes require version-specific evaluation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "bf92d966e3b5041cdfa347acd8f652c0743db28b88e83346afc4c94e079cedcb", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Knowledge graph entity alignment", "url": "https://docs.typesafe.ai/cookbooks/entity_alignment", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official worked example aligning entities through typed decision questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c5aa6dba0b497d9b95dcfaad4381501ab0c0a260a254481be50b80d5996d5145", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Lies, Damned Lies, and Benchmarks - TypeSafe AI Blog", "url": "https://typesafe.ai/blog/antibenchmaxxing", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "TypeSafe discussion of benchmark incentives and evaluation practice; a vendor viewpoint.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "498eb3997a4dffb4c1d627f1b62bb8169eb9a97ab21f760949e00e1ed700c8e5", "retrieved_at": "2026-10-16T10:07:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Line-by-line search", "url": "https://docs.typesafe.ai/cookbooks/semantic_find", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official semantic-finding example with decision questions over candidate text.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "92b23c001081fd4244a99e7162e87f8152306865c2906af8922fcc29587171d6", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Parallel questions", "url": "https://docs.typesafe.ai/cookbooks/parallel_questions", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official example of sharing input state across multiple parallel decision questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1f4574474139321a1e05b31887a18ffe1dd3235169276aa4b600c6ebc110aee8", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Re-ranking", "url": "https://docs.typesafe.ai/cookbooks/rerank_typesafe", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official reranking workflow using bounded evaluation questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7d8969ab0d851fb0296d5e86c53eac03f3ba3dc0b779aecd7903604f51a6ce11", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Reading Right, Acting Wrong — author evaluation dashboard", "url": "https://yubol-bobo.github.io/System_one_model_evaluation", "category": "Guides and evaluations", "topic": "paper-evaluation-dashboard", "description": "Author dashboard presents paired factual-probe and action-decision diagnostics across agent tasks; probe correctness does not establish complete understanding or isolate the cause of a wrong action.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e3a51c32cd3aafadd488d2bfe257027e4b649975d515359da0b14fb5556e9529", "retrieved_at": "2026-10-16T12:44:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SDE cascade", "url": "https://docs.typesafe.ai/cookbooks/sde_cascade", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official structured-data-extraction cascade that checks inexpensive model output with Jev before escalating to a reasoning model.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "90ab0294cb0d83b0ece3343edf11c6d492ce0236f4e21b434bbdbcb7c5500b74", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Self-consistency: choices", "url": "https://docs.typesafe.ai/cookbooks/consistency_choice_cookbook", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official procedure for inspecting consistency of categorical Choice decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a1a8481c676fd4b6419a9f26feeacd1f223624cb17f438ac8b640597bbc03900", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Self-consistency: nouls", "url": "https://docs.typesafe.ai/cookbooks/consistency_noul_cookbook", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official procedure for inspecting consistency of Noul yes/no decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7d0a90fc22b71703d4dc38a901032afbe5a22bb911134d553654f143a9a6fd41", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Simple memory system", "url": "https://docs.typesafe.ai/cookbooks/basic_memory", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official example of using decision primitives to manage application memory.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8e894e5b93ac45ffbe5b18997e9f4e811c99baf9f4504a894fa816df276f41ef", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Skill suggestion", "url": "https://docs.typesafe.ai/cookbooks/skill_suggestion", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official example of selecting relevant skills from explicit candidates.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c59bf591db3d45e20b90a93a4526775e5188c479c6366b9a4267cb37ab51fc90", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Smart home assistant demo", "url": "https://docs.typesafe.ai/demos/smart-home", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official smart-home demo using typed decisions to route actions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ec928346966d7634de65e1877cec1b5741d1a0319fdfc3a6ea99641b408cbc5a", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Speculative fan-out", "url": "https://docs.typesafe.ai/patterns/fan-out", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official fan-out pattern for evaluating multiple questions or candidate actions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7a81fc0647792c2e2b13afef2c044e80756de16161cd0840473ee46276a06363", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "State", "url": "https://docs.typesafe.ai/concepts/state", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official explanation of the shared input state supplied to decision questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0058f0fc9c583183912061fc57b583ec51a5fc354e0163d01b276d488e98a8d9", "retrieved_at": "2026-10-16T10:07:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Structure recovery", "url": "https://docs.typesafe.ai/cookbooks/autoformat", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official two-request method for restoring Markdown structure from unformatted text using line-continuation and block-type decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "cb5db3e079d337d0b5923e1d122a9c65612e342061d42939ed1aea130f3a58ba", "retrieved_at": "2026-10-16T10:07:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "System One", "url": "https://docs.typesafe.ai/concepts/system-one", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official conceptual guide to the System One decision interface; vendor framing, not independently established cognitive equivalence.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f3ef88c39b3e6a52db708bef199a8d72be5d4f1ae3b58e256d1fae141f1836c2", "retrieved_at": "2026-10-16T10:07:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Twitch chat filter", "url": "https://docs.typesafe.ai/cookbooks/twitch_chat_filter", "category": "Guides and evaluations", "topic": "official-decision-guide", "description": "Official chat-filtering example; application policy and error-cost choices remain separate.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5cb7a3d1d62eafccc48a0962f1fa9a2f3f4f6c6b5acac17b7ec51ff01d99d8fb", "retrieved_at": "2026-10-16T10:07:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Typed Decision Leaderboard — JEV & open-Jev answer verifiers, one identical test set", "url": "https://huggingface.co/spaces/mayafree/typed-decision-leaderboard", "category": "Guides and evaluations", "topic": "evaluation-interface", "description": "Independent typed-decision verifier leaderboard on a shared labeled set with published grading materials.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "47e39288c7db9163f574130dfdcaf4eed6a78144f935ab9e534f9939f2709174", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "TypeSafe Jev: 5x Faster, 25x Cheaper, 98.5% Correct | Entagl", "url": "https://entagl.com/blog/typesafe-jev-benchmark-ai-decision-models", "category": "Guides and evaluations", "topic": "decision-model-evaluation", "description": "Author-run Jev comparison on labeled decisions and customer-workflow replays; workload-specific reported findings, not reproduced here.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "13982f26419fc6ab5a43a5467b9bf1982e858d32452a8651b1d28a34a0aa99ea", "retrieved_at": "2026-10-16T10:07:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ZTC Gate Tetris", "url": "https://huggingface.co/spaces/FINAL-Bench/Tetris-JEV-LAYA-ZTC", "category": "Guides and evaluations", "topic": "evaluation-interface", "description": "Tetris gating simulation driven by publisher-measured judge statistics; not a record of the models actually judging Tetris moves.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "15f8b252cfaca3e97064a7bb183db9f1a5109511589045c9c90de91fb999b7eb", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "A decision model that learned to cook and cooperate", "url": "https://huggingface.co/rodneyslafuente/decision-model-rl-overcooked", "category": "Models", "topic": "decision-model-card", "description": "Overcooked cooperative-game decision policy, with discrete native-action scoring and shared-policy evaluation artifacts.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e0b9582b04bffa576ca3cbc8a4a7793d9cea2dafa66459b7737613237e77bfb6", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "aimeigaoshou/agent-jev", "url": "https://huggingface.co/aimeigaoshou/agent-jev", "category": "Models", "topic": "decision-model-card", "description": "AgentJev model card documenting a compact state-to-distribution decision interface without free-text decoding.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6d815be329e11626d89c9d41271010052b106fefd3b3d571f8dcd7b92552ce41", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "allenai/truthfulqa-truth-judge-llama2-7B", "url": "https://huggingface.co/allenai/truthfulqa-truth-judge-llama2-7B", "category": "Models", "topic": "decision-model-card", "description": "Open TruthfulQA truthfulness judge checkpoint, relevant to comparing model-derived evaluation labels.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "94bc5deeb52a357d3a7223c8901ceecb829a271cce7220f6c3f36b88b478b48a", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Archai JEV Nesso 4B Ita (System One Decision Engine)", "url": "https://huggingface.co/nickprock/archai-jev-nesso-4b-ita", "category": "Models", "topic": "decision-model-card", "description": "Italian candidate-token decision checkpoint covering moderation and intent-routing tasks; capabilities remain publisher claims.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "428b7bb08a0c39417ab29922bb7703c9742d79ff831ae0d996f3c87d51209258", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "bit-jev-2b-distilled", "url": "https://huggingface.co/jinghao1632/bit-jev-2b-distilled", "category": "Models", "topic": "decision-model-card", "description": "Bit-Jev distilled model card with bilingual typed-question interfaces and local runtime instructions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "77a2c417a4d35ff72c133456810af3f144d12745565bd65ee87c57282bdb8ef2", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Bonsai-JEV", "url": "https://huggingface.co/dakebakery/Bonsai-JEV", "category": "Models", "topic": "decision-model-card", "description": "Bonsai-based image-capable decision model documenting soft-label distillation and a decision head.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c99483ff5d2860705f4dd0141f209c8900f18545dee78e3e990054d9c3e11468", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Cached causal typed scorer (arm B, Jev-like reproduction)", "url": "https://huggingface.co/abidlabs/jev-typed-decisions-causal-0.6b", "category": "Models", "topic": "decision-model-card", "description": "LoRA checkpoint reproducing the cached causal scoring arm of a public typed-decision experiment.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b5075891650ae909a27ead37cbf55bf07c27e001df7f2410b67b24fe8b6c4fb6", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "codeguard-jev-style-1.5b", "url": "https://huggingface.co/JackKozmo29/codeguard-jev-style-1.5b", "category": "Models", "topic": "decision-model-card", "description": "Code-review triage model returning structured verdicts, risk levels and flags for pull-request diffs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "23c92b5560614ac4e00a093364808d0993f337ad1db9c5aa3b7d4776ecb4cef6", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CUA-S1 Forge RLCD-v3", "url": "https://huggingface.co/NewSonnet/cua-s1-forge-rlcd-v3", "category": "Models", "topic": "decision-model-card", "description": "Verifier-guided training experiment for one-step computer-use form actions on synthetic held-out form signatures.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4ffd93f825a1562fbc50797ce226a2b84048d93e0b33abf9039d6f9dfb1aaabc", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Decision head for Qwen3.5-4B — RLCD, 32K one-pass", "url": "https://huggingface.co/AstroHan/decision-head-qwen3.5-4b-rlcd-32k", "category": "Models", "topic": "decision-model-card", "description": "Controlled decision-head experiment varying data coverage while fixing the training recipe, with held-out transfer evaluations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7691c85a0aae330536b170e9729f99f7fd5c0f8a3bdbb8e0c982b6523547cb43", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "diffcider-typed-decisions", "url": "https://huggingface.co/sgaseretto/diffcider-typed-decisions", "category": "Models", "topic": "decision-model-card", "description": "Diffusion-model adaptation reading Yes/No logits for typed Choice, Score and Noul decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4aefd13cb55286bc81976c01bcca7d9add47af56b6cfda9f4e2d017461337f2d", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "EmbeddingGemma 2 · typed decisions", "url": "https://huggingface.co/mrm8488/embeddinggemma-2-ft-typed-decision", "category": "Models", "topic": "decision-model-card", "description": "EmbeddingGemma adaptation for typed questions over support, invoice, security and agent-trace workflows.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "bd0545657a6878f437b9fa9989251bb53bce401808dc5ea82215ae4caefc5772", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "exu-qwen3.5-2b-typed-decisions", "url": "https://huggingface.co/Ruivalim/exu-qwen3.5-2b-typed-decisions", "category": "Models", "topic": "decision-model-card", "description": "Exu decision head for agent and workflow traces, with option scoring and fitted-temperature probabilities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1c3e9105015f1f3d1b9835bc17e93590a67d26f9712e0517cead906cbea6ce87", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Granite-20B-FunctionCalling", "url": "https://huggingface.co/ibm-granite/granite-20b-functioncalling", "category": "Models", "topic": "decision-model-card", "description": "Granite function-calling checkpoint covering nested, chained and parallel tool calls as a generative comparison baseline.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "808aae20be1bfd0e49017b0cafe23f6e92e1d161b6794420dd22760cf88cf2d3", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jeb-typed-decisions", "url": "https://huggingface.co/IJyad/jeb-typed-decisions", "category": "Models", "topic": "decision-model-card", "description": "Arabic operational-workflow decision checkpoint covering tickets, invoices, alerts and agent traces.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "bdabfdccb0b41e11b1b9cbff7a7cff733affbe55fc9f68d149910c69fa7fac73", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jeva-4B — author checkpoint package", "url": "https://huggingface.co/LIQIIIII/Jeva-4B", "category": "Models", "topic": "decision-model-card", "description": "Author checkpoint bundle for Jeva with backbone, adapter and decision-head files; weights were not loaded, and the inspected card specifies no model license.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "04108705770dea7fe225995debbbd7727c58112a40afd53dfb1b110afd55a23f", "retrieved_at": "2026-10-16T12:45:32.714336+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevEmbed-KaLM-Embedding-V2.5", "url": "https://huggingface.co/HIT-TMG/JevEmbed-KaLM-Embedding-V2.5", "category": "Models", "topic": "decision-model-card", "description": "JevEmbed embedding checkpoint with documented Choice, Score and Noul scoring over its training corpus.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8efb6af5042af2fa1a5fa997da5175d8ef973c685bf5a4a829c40ed393e1881c", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jevons-lfm25-1.2b-systemone", "url": "https://huggingface.co/gopalanj/jevons-lfm25-1.2b-systemone", "category": "Models", "topic": "decision-model-card", "description": "Jevons seed adapter for option-token scoring, with typed outputs assembled by the serving implementation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "12edbb54385ca2268b5515f87b384125d6652d5c13ccdfa98a0bfd07ef6b53df", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JudgeLM Model Card", "url": "https://huggingface.co/BAAI/JudgeLM-33B-v1.0", "category": "Models", "topic": "decision-model-card", "description": "Vicuna-based JudgeLM checkpoint for evaluating open-ended answer pairs; useful as a generative judge comparison.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "c82aa6f432dc73e461607e54dff1a02a80c652b18fa8c17f60f9176186f7260a", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "katanemo/Arch-Router-1.5B", "url": "https://huggingface.co/katanemo/Arch-Router-1.5B", "category": "Models", "topic": "decision-model-card", "description": "Arch-Router checkpoint addressing preference-aware LLM routing rather than general answer generation.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "3b2daa5c898626e2b4d1279d57cad8cb61dcb850a4af8aabac032de2ce70a7dd", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "kenning-xl-v0.6 (experimental)", "url": "https://huggingface.co/systemonedev/kenning-xl-v0.6", "category": "Models", "topic": "decision-model-card", "description": "Experimental Kenning decoder variant using constrained answer-token readout behind the System One interface.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "eedd2b26d8ac00fc994ff7b381660aeaa436a27910743668cd090c5fe9da3f37", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Laya typed-decisions, fine-tuned on browser decisions", "url": "https://huggingface.co/AIMultiple/laya-typed-decisions-browser-ft", "category": "Models", "topic": "decision-model-card", "description": "Browser-agent fine-tuning experiment for Laya, including reported negative results and reproducibility details.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0f16e2116d6ac55b8adb717baf79a7b9be9ae75d1014b2cfaf9ecf24f6159d81", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "LFM2.5-Encoder-350-Prompt-Router", "url": "https://huggingface.co/LiquidAI/LFM2.5-Encoder-350M-Prompt-Router", "category": "Models", "topic": "decision-model-card", "description": "Encoder-based prompt router that scores user-defined routing lanes in one pass, an adjacent routing baseline.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "4562b45a5d4b4035c734408493d9a79da6cbf88e10a865666620c2d8e8d4d92e", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "LightJev 0.6B — Typed Decisions v0.2", "url": "https://huggingface.co/rongxinzy/LightJev-0.6B-typed-decisions", "category": "Models", "topic": "decision-model-card", "description": "LightJev candidate-scoring checkpoint for finite-answer English workflow questions, with intended-use boundaries.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8e1c551fe3e2e67aabad650dc2cf8140e44e5f3f1371e9141baf12e1a69d84fb", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "llm-jp-4.1-33b-systemone-experimental", "url": "https://huggingface.co/shunmoridev/llm-jp-4.1-33b-systemone-experimental", "category": "Models", "topic": "decision-model-card", "description": "Experimental Japanese LLM-jp candidate-token scorer; the card explicitly disclaims demonstrated general Jev-equivalent quality.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1f367bda8cdc67f0b37b49cf3c79b4348bdb6393e14f197ed2bb8d8e10e2deed", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "local-system-one-student", "url": "https://huggingface.co/Mannedood/local-system-one-student", "category": "Models", "topic": "decision-model-card", "description": "Distilled encoder for one fixed coding-agent action schema; the card explicitly limits claims of Jev equivalence.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "188f6854f477fddf9c83a15063610cbddf802648b33faf709c541d93ce9eb8df", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "MD-Judge for Salad-Bench", "url": "https://huggingface.co/OpenSafetyLab/MD-Judge-v0.1", "category": "Models", "topic": "decision-model-card", "description": "MD-Judge safety classifier for question-answer pairs, offering an adjacent guardrail and evaluator baseline.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "e4d28655a09f2b81da6fb8306f99591e4ea71dd786b290a347fa14c5e07cd2c4", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "mJev-Qwen3-VL-4B-RLCD", "url": "https://huggingface.co/SoMarkAI/mJev-Qwen3-VL-4B-RLCD", "category": "Models", "topic": "decision-model-card", "description": "Multimodal mJev candidate-selection checkpoint with visual context, probability outputs and documented runtime.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "24bfcae47cdedd8f4b9f27265110cd24a2eba8e5c32ab964d562c17901bee87b", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Modality Router (Merged) - Smart Output Modality Selection", "url": "https://huggingface.co/vllm-sr/mmbert32k-modality-router-merged", "category": "Models", "topic": "decision-model-card", "description": "ModernBERT modality router choosing output modalities for vLLM Semantic Router, an adjacent request-routing artifact.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "6e140229eed4439b6439090628c6123296b524d3260f212357c06b4ee5083b16", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "modernbert-ja-310m-jev", "url": "https://huggingface.co/argos1111/modernbert-ja-310m-jev", "category": "Models", "topic": "decision-model-card", "description": "Japanese ModernBERT cross-encoder for scoring state/question/candidate pairs through a Jev-compatible interface.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fc2902bdae110c0ce565721a53f3076c587c47a193fd75e28c25d0ae55bfe363", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Nandi-Mini-150M-Tool-Calling", "url": "https://huggingface.co/FrontiersMind/Nandi-Mini-150M-Tool-Calling", "category": "Models", "topic": "decision-model-card", "description": "Compact Nandi model card describing single-turn tool-call generation, an adjacent alternative to bounded action selection.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "ef871516175a61b7ed20c5d119e163a21f248d2f0922d87c361ceaebc070ee36", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Nano-Jev v1.0", "url": "https://huggingface.co/sdmlai/nano-jev", "category": "Models", "topic": "decision-model-card", "description": "Nano-Jev small cross-encoder decision model for RAG pipelines, returning probabilities over supplied candidates.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d69ddc4f0b0f384c329a1553ffebb9752dd5973483d9a9a3947b32e5abf5eb48", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "od1-typed-decisions · Open Decider (4B specialist)", "url": "https://huggingface.co/mvbalaji/od1-typed-decisions", "category": "Models", "topic": "decision-model-card", "description": "Open Decider specialist checkpoint with an explicit not-answerable outcome alongside typed answer probabilities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "52f311d79ae89327b8c1790f9f1bd3418bb3b859b97c3ef57b6aee2129c6648d", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "open-jev-deberta-v3-large", "url": "https://huggingface.co/com-kotobalabs/open-jev-deberta-v3-large", "category": "Models", "topic": "decision-model-card", "description": "Independent DeBERTa implementation of typed questions and candidate-probability readout, without claiming original Jev weights.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "885eff8a574bb85692da85c550a86742406e6a65b6825d1476db524a65c0b5f9", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "OpenThai-SystemOne", "url": "https://huggingface.co/iapp/OpenThai-SystemOne", "category": "Models", "topic": "decision-model-card", "description": "Thai-English decision model card describing bounded choice, ordered-score and yes/no probability interfaces.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0413faacfc1dfcddd48929bc9846ae3e2d408b8753d6e24333438eb4400a5bf4", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "projectj/Flight-Autopilot-Decision-Model-Maverick1.0", "url": "https://huggingface.co/projectj/Flight-Autopilot-Decision-Model-Maverick1.0", "category": "Models", "topic": "decision-model-card", "description": "Flight-simulator decision-model project with documented action-selection experiments; not evidence of operational aviation suitability.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f6964130bfb7c3f1d7301c5656908bb07465b752b501325d065b328b9b19ec7c", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Qwen3-0.6B RLCD decision model", "url": "https://huggingface.co/anthonym21/qwen3-0.6b-rlcd-decision", "category": "Models", "topic": "decision-model-card", "description": "Typed decision checkpoint documenting supervised warmup followed by bandit-feedback RLCD and option-distribution outputs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ed56eb33c022345184763391c1acac244b018aae124b4e5c70ab41eb8e78dfb8", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Qwen3-14B RLCD Decision LoRA", "url": "https://huggingface.co/Foodoo1/Qwen3-14B-RLCD-Decision-LoRA", "category": "Models", "topic": "decision-model-card", "description": "Qwen adapter for parallel single-token scoring of several schema fields, with candidate-token readout documented.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fe09e025cc078e15938eac48fe16e46c9fc4f17d3dd45c21246a510114825e48", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "raz-systemone-nli-xsmall (v5)", "url": "https://huggingface.co/RazvanManolache/raz-systemone-nli-xsmall", "category": "Models", "topic": "decision-model-card", "description": "NLI cross-encoder that maps state/candidate entailment scores to typed answer distributions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "de7ce796fe54600a0fb580471886ff727ff73540fca67ea56a28f6ce2a67ad5b", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "rubert-tiny-turbo-typed-decisions-ru", "url": "https://huggingface.co/yyhlm/rubert-tiny-turbo-typed-decisions-ru", "category": "Models", "topic": "decision-model-card", "description": "Russian specialist decision checkpoint limited to the translated benchmark's workflow schemas, with distributions for each answer.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b1b73d10bace50dde5d642721ed849a905702cd62abf9e815e6e15bad0b7dd0b", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "s1-llm-auto-router", "url": "https://huggingface.co/system1models/s1-llm-auto-router", "category": "Models", "topic": "decision-model-card", "description": "CPU request classifier predicting routing signals such as task category, difficulty and stakes for an LLM router.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "280ab53b776e0cea9ade7d85e07fda01a0a0bc04be324a6557b9c70bc1941f1b", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Schema-conditioned candidate scorer (DeBERTa-v3-large)", "url": "https://huggingface.co/mobarmg/jev-schema-scorer-deberta-v3-large", "category": "Models", "topic": "decision-model-card", "description": "Schema-conditioned DeBERTa scalar scorer with deterministic grouping into Choice, Noul and Score answers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3a64efe5e894ef018cfe888dc8118479f7e643d52f9b073870772d099466b99b", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SourceHat RLCD: typed decisions with Qwen3.5-9B", "url": "https://huggingface.co/SourceHat-Labs/RLCD-Qwen3.5-9B-Gated-Decision", "category": "Models", "topic": "decision-model-card", "description": "Gated decision-head study distinguishing commitment, alternatives and abstention over supplied choices.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "08c4c7ebc4f838a33e22a09f54e33167b72c5cc2693b43490e320d5874f96dfc", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "StartLux-Decision-v1.1-27B-Instant", "url": "https://huggingface.co/startlux-models/StartLux-Decision-v1.1-27B-Instant", "category": "Models", "topic": "decision-model-card", "description": "Multimodal StartLux decision checkpoint with typed questions, bounded image inputs and candidate probabilities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "807aa5311a50cbc8f3f8fcf1a8deeb289d3b54354fb11c0bfc1f0a24e121fd85", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "System One (distilled)", "url": "https://huggingface.co/shreyanbr/system-one-distilled", "category": "Models", "topic": "decision-model-card", "description": "Distilled DeBERTa model using teacher-generated supervision for typed routing, classification and gating questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ce05d912982b284d0dc2a12b3b32b77deebfb635c37e4eb0cb973451cda130a0", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "System-One QLoRA", "url": "https://huggingface.co/yocoms/system1-qlora", "category": "Models", "topic": "decision-model-card", "description": "QLoRA single-pass decision-scorer experiment with held-out calibration reporting and explicit independent-project scope.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0c3099ae4ce0ab88441948f9add646de818d3b1dcbe3470bd35592a16575e99a", "retrieved_at": "2026-10-16T10:05:30+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "systemone-lite-0.5b", "url": "https://huggingface.co/dwidlee/systemone-lite-0.5b", "category": "Models", "topic": "decision-model-card", "description": "Small local System One-compatible model using option-restricted next-token scoring and documented synthetic training splits.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f9885507135f8c927d2a0045be3d06ad198f3d56e544ef328c61d5a4d0f2f4bc", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Technique Router (MiniLM)", "url": "https://huggingface.co/chopratejas/technique-router", "category": "Models", "topic": "decision-model-card", "description": "MiniLM router selecting image-compression techniques from natural-language queries, an adjacent task-specific routing example.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "f55c34e285f414fde79774660344e9b6eb27e792765b93338ded8e5a660f445e", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Tiny-Jev", "url": "https://huggingface.co/lostargon/Tiny-Jev", "category": "Models", "topic": "decision-model-card", "description": "Independent Tiny-Jev implementation documenting typed probabilities and its separation from TypeSafe's weights and code.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "72ea6911285f5f5d63e4acff5b88d9632e22d7ca79d6d9c99c2e31e4f0e10697", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "typed-decisions-minilm-l6-specialist", "url": "https://huggingface.co/adaptive-classifier/typed-decisions-minilm-l6-specialist", "category": "Models", "topic": "decision-model-card", "description": "Workflow-specific classifier heads over a frozen MiniLM encoder; the card explicitly identifies supervised specialist scope.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "52c5e2356ef7057d263fd4561abf556b885d4033d314b7f8b919e85db8e8e836", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ufakzeka-karar — Turkish typed-decision checkpoint", "url": "https://huggingface.co/ufakai/ufakzeka-karar", "category": "Models", "topic": "decision-model-card", "description": "Turkish typed-decision checkpoint with shipped temperature scaling; its card flags test-informed training and worsening support-task calibration. Weights were not loaded or independently evaluated.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2eb35606e2c919b4683242ddc4d1f28e0bae6e0bf844559db771c4f58368d2f5", "retrieved_at": "2026-10-16T13:16:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "vistalab-system-one-12b", "url": "https://huggingface.co/UlkuTuncerKucuktas/vistalab-system-one-12b", "category": "Models", "topic": "decision-model-card", "description": "Turkish System One checkpoint using candidate-letter probabilities for finite-option questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "61feadcf94d6444698a59d888b08ba1b6f2cc09161b7c164fba77280e6395c07", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "W1-JEV", "url": "https://huggingface.co/WhaletechAI/W1-JEV", "category": "Models", "topic": "decision-model-card", "description": "W1 diffusion-model card documenting structured decision tasks and comparison artifacts, with publisher-reported performance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9c8d525f3c497b1ae2e236b7802b75f5c778aa2f51979c5e7fd8778a0feb0db0", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Yway (ရွေး)", "url": "https://huggingface.co/aungthuhein-dev/yway-system1-xlmr", "category": "Models", "topic": "decision-model-card", "description": "Burmese XLM-RoBERTa decision model with Choice, Noul and Score questions and per-answer probabilities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3968b25c463f4b787d67b12b965020934ac7cf2fa03effafcc1063d03aad1575", "retrieved_at": "2026-10-16T10:05:24+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Beyond Answer Confidence: A Controlled Audit of Self-Knowledge in a Black-Box Decision Model", "url": "https://arxiv.org/abs/2610.01006", "category": "Papers", "topic": "direct-decision-research", "description": "Controls information and surface cues to distinguish Jev answer confidence from self-knowledge; black-box behavior does not establish access to an internal knowledge state.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "abee899ad2ed5e5331283d34002c1129e09706fbf39fc8aab1aee6deac690882", "retrieved_at": "2026-10-16T11:20:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Bongard: Training Machine Intuition", "url": "https://arxiv.org/abs/2609.39111", "category": "Papers", "topic": "direct-decision-research", "description": "Shared evidence encoding and candidate-probability readout, trained with representation and action-outcome feedback. Results remain author-reported.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "eb8f13a8b534af9a3b9bcf907d3948c81f5bab91e0b9f3d80c26e7fd97299c00", "retrieved_at": "2026-10-16T10:01:30+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibrated to Whom? Persona and Language Effects on Cultural Values in JEV", "url": "https://arxiv.org/abs/2609.36399", "category": "Papers", "topic": "direct-decision-research", "description": "English/Arabic persona-conditioned Jev survey audit; the effects concern simulated respondents in one model, not measurements of real populations.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "252516d8d2459e2ab4eb43e7407cdf4d4e7d2c27a420ce9985e5d1395032d75d", "retrieved_at": "2026-10-16T11:20:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Can Jev Judge Radiology Reports? Evaluating a System One Model for Clinical Factuality", "url": "https://arxiv.org/abs/2609.27607", "category": "Papers", "topic": "direct-decision-research", "description": "Checks statement support in both directions between generated and reference radiology reports; agreement depends on report size and error scope, and a specialized comparator remains stronger on clinically significant errors.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "3239e5a740f99121cc1ba98c04e31709b3a18c695b6860a8520f838721714e79", "retrieved_at": "2026-10-16T12:16:34.052101+00:00", "ledger": "research-sources.jsonl" }, { "title": "Decision Hijacking: Prompt Injection Attacks on Jev's Typed Probabilistic Decisions", "url": "https://arxiv.org/abs/2609.28613", "category": "Papers", "topic": "direct-decision-research", "description": "Jev injection audit separates probability shifts from attacker-target selections; adaptive attacks are checked on fresh validation calls over reconstructed InjecAgent cases.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "898b72728a465d950ea191589cd6c4aebd1858bc658315ccf6cb7255a62a0d40", "retrieved_at": "2026-10-16T11:40:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "Evaluating Decision Models for Text Annotation in Computational Social Science", "url": "https://arxiv.org/abs/2609.24574", "category": "Papers", "topic": "direct-decision-research", "description": "Evaluates decision models and confidence-routed LLM cascades on selected zero-shot social-science annotation tasks; lower cost coexists with task-dependent errors and overconfidence.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "0f5170a5e8987b1e9e6f432e25e6decd3e18e1353f489d024310b7afc116cff3", "retrieved_at": "2026-10-16T11:40:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "JEV as a Judge for Agent Trace Security: An Empirical Comparison with Generative LLM Judges", "url": "https://arxiv.org/abs/2609.34862", "category": "Papers", "topic": "direct-decision-research", "description": "Retrospective agent-trace risk classification compares Jev with generative judges across four collections; rankings vary by dataset, and latency/cost summarize successful or valid calls.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "4aee8edd65542ffc0ba8800c681db222984cbf1bc40b239c829feb823aab3ffa", "retrieved_at": "2026-10-16T11:40:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "Jev for Scientific Decisions: Evaluating Semantic Choices and Their Consequences", "url": "https://arxiv.org/abs/2609.24965", "category": "Papers", "topic": "direct-decision-research", "description": "A small scientific-workflow study separates semantic Choices, computed quantities and final labels; selection errors in comparison models can change intermediate counts without changing the final label.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "3c4bd5724102faa33a0da01e5fdee7b3e9d0ec4eb2533b02de24e85409147f4d", "retrieved_at": "2026-10-16T11:52:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Jev in Medicine: A Benchmark Evaluation", "url": "https://arxiv.org/abs/2609.34024", "category": "Papers", "topic": "direct-decision-research", "description": "Preliminary medical-benchmark evaluation of Jev 1.13 reports task-dependent accuracy, calibration and weak unanswerable-question detection; it does not establish clinical deployment validity.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "25988c14a48518f0e90f24f2f9b907edc4a4a96ef2f51fef7735cf38e89a5209", "retrieved_at": "2026-10-16T11:52:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Jev-Mobile: Jev as an Executor for Mobile GUI Agents", "url": "https://arxiv.org/abs/2609.30186", "category": "Papers", "topic": "direct-decision-research", "description": "VLM local-goal planning with Jev action selection over mobile accessibility trees; time and cost comparisons are conditioned on successful trajectories.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "b387c1cfe77f6e06477415944b3923189a1f0ca30be0680f01815cb65b0067bc", "retrieved_at": "2026-10-16T10:59:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "JEV-Star: Fast, Low-Cost StarCraft II Control with Language-Model Planning", "url": "https://arxiv.org/abs/2609.27331", "category": "Papers", "topic": "direct-decision-research", "description": "Jev action selection paired with language-model planning for StarCraft II; concurrent interface changes prevent isolating the planning effect.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "1200b29815b7e13b36740843f0286ad6316d575dfa61d30989bd4355d7e0ffb0", "retrieved_at": "2026-10-16T10:59:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "JevOut: Natural Context Can Flip Decision Models", "url": "https://arxiv.org/abs/2609.30243", "category": "Papers", "topic": "direct-decision-research", "description": "Probability-guided context additions redirect initially correct decisions in four systems while preserving questions and options; human validity checks cover only a sampled subset of successful attacks.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "3429d7a57eb41ced8ce4f3b9ca20f201017d85e97fd65d2e818db23bb354f0a3", "retrieved_at": "2026-10-16T11:40:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "Judges Follow the Reference: The Derivation Gap in Verifying Tool-Call Arguments", "url": "https://arxiv.org/abs/2610.18497", "category": "Papers", "topic": "direct-decision-research", "description": "Tool-argument judges including Jev reject some correct derived values without visible references; model-generated references are evaluated on three benchmarks.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "503058658264a84669534e672294ba2af62736c2ebd3ec95d08d1414d7e215a4", "retrieved_at": "2026-10-16T11:20:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "KITE: Scaling Jev Population Experiments with Sparse Flagship Calibration", "url": "https://arxiv.org/abs/2609.27535", "category": "Papers", "topic": "direct-decision-research", "description": "Combines Jev state tables, sparse flagship-model anchors and human–model discrepancy estimates; scaling a simulated population does not increase the underlying human evidence.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "7a1c1f4bb28c771a5da44f49801a8547590f941e331ec8bf081f6fcee169214e", "retrieved_at": "2026-10-16T11:52:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Mnemon: Raw Records, Fast Judgments, Slow Thoughts", "url": "https://arxiv.org/abs/2609.36059", "category": "Papers", "topic": "direct-decision-research", "description": "Keeps raw memory records, separating Jev relevance judgments from LLM search planning and answer generation. Results remain author-reported.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "3d090d44f7c2ae808bf730adb6279d42462fd250ad70f2a6c4d58ab6798abbbc", "retrieved_at": "2026-10-16T10:01:29+00:00", "ledger": "research-sources.jsonl" }, { "title": "More Choices, Fewer Decisions: Ordinal-Scale Bias in JEV-like Direct-Decision Models", "url": "https://arxiv.org/abs/2609.38827", "category": "Papers", "topic": "direct-decision-research", "description": "Studies underuse of ordinal labels in direct-decision models; finer scales intensify compression, while better accuracy or shuffled option order does not necessarily restore scale utilization.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "11da327e2f623bfeb039d1d0f75b174b635d6dcb91612ae0fe8bf5f6ca47db7e", "retrieved_at": "2026-10-16T12:16:34.055084+00:00", "ledger": "research-sources.jsonl" }, { "title": "One Word Opens the Gate: The Option-Channel Attack on Typed Decision Models as Agent Guardrails", "url": "https://arxiv.org/abs/2610.12292", "category": "Papers", "topic": "direct-decision-research", "description": "Tests option-channel attacks on seven open-weight models; reversed decisions can remain confident. The reported experiments do not test hosted Jev.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "b2270f7af20a7bd8da9ca49784aa4c2145cd500c8d92f194d838ac3334755ff6", "retrieved_at": "2026-10-16T11:20:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "REFLEX with Jev for Efficient Selective Control in LLM Agents", "url": "https://arxiv.org/abs/2609.26532", "category": "Papers", "topic": "direct-decision-research", "description": "Confidence-gated Jev control with LLM fallback; gains depend on action alternatives and baseline routing. Results remain author-reported.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "f6cbac19c9acb5fb9e0d10e8afb293704c90acb01ad1084fd5cc2ba182bbe968", "retrieved_at": "2026-10-16T10:01:25+00:00", "ledger": "research-sources.jsonl" }, { "title": "Same Scores, Different Decisions: Evaluating JEV and Language Models for Legal Document Understanding", "url": "https://arxiv.org/abs/2609.27678", "category": "Papers", "topic": "direct-decision-research", "description": "Repeated ContractNLI evaluations show that opposing corrections and regressions can leave aggregate accuracy unchanged; consistent answers may remain wrong, while majority-vote gains are small and uncertain.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "e77c66ad5c4cc7007cadcf394cc654b163cbfd376127101c97c8710157dd8b4e", "retrieved_at": "2026-10-16T11:52:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Same-Number Citation Swaps: Stress-Testing Jev as a Financial Evidence Judge", "url": "https://arxiv.org/abs/2610.08675", "category": "Papers", "topic": "direct-decision-research", "description": "Controlled same-number citation swaps test Jev source verification; detecting wrong financial roles trades off against accepting equivalent evidence.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "a71eeed52135068e8db08487008090e5f61b830cd1632f2c8a827f8c569d2366", "retrieved_at": "2026-10-16T10:59:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "Selective Reasoning for Typed Decision Models: When to Start, Read, and Stop", "url": "https://arxiv.org/abs/2610.19644", "category": "Papers", "topic": "direct-decision-research", "description": "Candidate-answer readout during interruptible reasoning, with gates for starting, reading and stopping; results depend on a fixed backbone and task suite.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "fc2f52f0f0bd8395c70abfe3511f2fe9c16e98c0108bd378199458aff6165968", "retrieved_at": "2026-10-16T10:59:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Specialized Decision Models vs. General-Purpose LLMs: Benchmarking Jev Across Knowledge, Reasoning, and Multilingual Tasks", "url": "https://arxiv.org/abs/2610.11978", "category": "Papers", "topic": "direct-decision-research", "description": "Knowledge, reasoning and multilingual multiple-choice evaluation of Jev; reported rankings and mathematical weaknesses apply to the selected models and benchmarks.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "a641b55cb05c921f04724be7efdc13c4749e88b1bd50dc932eb5b499c7e96b90", "retrieved_at": "2026-10-16T11:20:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Suggested Is Not Superseded: A Pre-Registered Evaluation of a System-One Model as a Memory-Retirement Gate", "url": "https://arxiv.org/abs/2610.15710", "category": "Papers", "topic": "direct-decision-research", "description": "Jev memory-retirement evaluation separates confirmed updates from suggestions; mixed LLM/human labels, shared errors and selection on a small development set constrain cascade claims.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "ee186f2c2b419284023f59137906108fb060f297c912b0e74766bb3f8e9272f7", "retrieved_at": "2026-10-16T11:20:08+00:00", "ledger": "research-sources.jsonl" }, { "title": "Token-Efficient Multi-Agent Collaboration via System One-Guided Computational Division of Labor", "url": "https://arxiv.org/abs/2610.08155", "category": "Papers", "topic": "direct-decision-research", "description": "System One controllers coordinate bounded multi-agent decisions while LLM workers reason; reported efficiency concerns the complete collaboration framework.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "2b2eba5bd11807c0cfa6b2acb758e2fb997e7f101dd8e6e81570ae51b7761084", "retrieved_at": "2026-10-16T10:59:20+00:00", "ledger": "research-sources.jsonl" }, { "title": "Typed Decision Models on Tabular Data", "url": "https://arxiv.org/abs/2610.19536", "category": "Papers", "topic": "direct-decision-research", "description": "Typed-decision classification and verification under content-preserving table rewrites; confidence can indicate instability without establishing correctness.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "a777f98a0cbaad39aa01cb1eaefa9fdf85c6301e49428577fd3b966ebf617bcb", "retrieved_at": "2026-10-16T10:59:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "VideoJev: System One Model for Video Understanding and Interaction", "url": "https://arxiv.org/abs/2610.18615", "category": "Papers", "topic": "direct-decision-research", "description": "Structured video decisions with temporal spans and candidate scores; cached and joint inference reuse visual computation. Results remain author-reported.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "5056cfae7d1c42b1545311a6e40fce0beaab00aefafcff3d8d08c247c65eecca", "retrieved_at": "2026-10-16T10:01:30+00:00", "ledger": "research-sources.jsonl" }, { "title": "Visual Jev: Accurate and Efficient Decisions from Shared Visual Context", "url": "https://arxiv.org/abs/2609.25845", "category": "Papers", "topic": "direct-decision-research", "description": "Shares an image prefix across batched forced-choice questions; reported training gains concentrate on represented task families, and warm amortized execution savings trade off against peak memory.", "relation": "direct", "verification_level": "abstract_checked", "content_sha256": "5d320d1459fc23efa2129117d8938563810fe99c4d40932bcea2cf6ae522ffad", "retrieved_at": "2026-10-16T12:16:33.987559+00:00", "ledger": "research-sources.jsonl" }, { "title": "0xjgv/system-one-router", "url": "https://github.com/0xjgv/system-one-router", "category": "Projects", "topic": "implementation-repository", "description": "Claude Code and Codex routing helper uses decision models to choose execution models and reasoning effort, with provider fallbacks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3fbe7a0ffbb1e98bf1c708759054c95294566eac923bdb44a4a76c8bc6fda0b8", "retrieved_at": "2026-10-16T10:37:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "0xLeonhack/tellagent", "url": "https://github.com/0xleonhack/tellagent", "category": "Projects", "topic": "implementation-repository", "description": "Market-research pipeline combining deterministic indicators, Jev gates and evidence-constrained language-model analysis.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "42899dfec38cf84e171a31a025927345ab1f3f75a3f25c6ee4e42e67751da253", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "aatifahmad123/jev-llm-routing", "url": "https://github.com/aatifahmad123/jev-llm-routing", "category": "Projects", "topic": "implementation-repository", "description": "RAG chatbot using Jev to route questions by complexity between local and hosted language models.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7fc3c9603b422b536aad9a3ee7f9cb9297e9aec5a877cbf461000658ab91f630", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "acbcore-labs/s1-core", "url": "https://github.com/acbcore-labs/s1-core", "category": "Projects", "topic": "implementation-repository", "description": "Python decision-model harness with typed question specifications, calibration gates, parallel requests and replayable decision logs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7a3990ef88bc02460ed66a45f5ab6f855e082f06db4d6d96cb876d9643b71f36", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "acquiredintelligence/attend", "url": "https://github.com/acquiredintelligence/attend", "category": "Projects", "topic": "implementation-repository", "description": "Event-stream coordinator uses Jev judgments to choose whether agent notifications should interrupt, queue, remain passive or be suppressed.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8f79d7a8b7eb2fd6a7b776c65fb75039ea4a5cbbd1de8316bcae29c42af67fff", "retrieved_at": "2026-10-16T10:37:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "adelvillar1/agnostic-router-kit", "url": "https://github.com/adelvillar1/agnostic-router-kit", "category": "Projects", "topic": "implementation-repository", "description": "Provider-independent workflow router with local-judge/Jev cascades, quota-aware model selection, execution controls and journaled runs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4b2daa63984fd582f52ad881c6b9f480b37da9306698d168d4a048292857a8e1", "retrieved_at": "2026-10-16T10:37:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "adelvillar1/dev-decisions", "url": "https://github.com/adelvillar1/dev-decisions", "category": "Projects", "topic": "implementation-repository", "description": "Git-workflow decision gates for sensitive-data scanning and diff classification, with recorded judgments for calibration.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "dd8fea193ba0206c0fa41caf3bbdee6a26191029c04d14f94bb10d3b8d88e295", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "adventists-ai/duplexjev", "url": "https://github.com/adventists-ai/duplexjev", "category": "Projects", "topic": "implementation-repository", "description": "Speech decision-model runtime and integrations for turn-taking, emotion and non-verbal audio judgments in voice agents.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "617b78b74034aea90e89cda8f9f9b5f21601831bdccc46422a59afd396893817", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "agent-labs-dev/fastbrowse", "url": "https://github.com/agent-labs-dev/fastbrowse", "category": "Projects", "topic": "implementation-repository", "description": "Browser agent separating bounded Jev action choices from language-model planning and source-backed answer composition.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "484652808893a9aa1fd9ea609f08c4e2b9c286be31807db2e51293e70fd5b035", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "AgentiLoop/Agent", "url": "https://github.com/agentiloop/agent", "category": "Projects", "topic": "implementation-repository", "description": "Mac automation agent with an optional Jev gate that reviews shell commands before execution.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "910bc47b5de12aa3f7db1ca70a8e83fb8ae9147ab184553ba06045c002fe7cab", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ai-ecoverse/kev.js", "url": "https://github.com/ai-ecoverse/kev.js", "category": "Projects", "topic": "implementation-repository", "description": "Browser runtime for Kev models using WebGPU or WASM, with tokenization and decision heads in a Web Worker.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "267ba2a98f769c0c209e5e90063fbe54c572360963710d680643ea472c3c49d9", "retrieved_at": "2026-10-16T10:37:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "aikeluargalee-tech/jev-v8", "url": "https://github.com/aikeluargalee-tech/jev-v8", "category": "Projects", "topic": "implementation-repository", "description": "On-demand cryptocurrency research terminal using Jev for structured decision support, without independently established investment performance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "73370114f2d00fc51ef45dd044d72587854ab441353a663f8923fad0b9ab9089", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ajayk/jev-go-sdk", "url": "https://github.com/ajayk/jev-go-sdk", "category": "Projects", "topic": "implementation-repository", "description": "Community Go client for the TypeSafe System One API, with typed requests and responses.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1b2d45770e31a64c2d325d892035de3aa8193a8facb66142ed97c37864485dce", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "andyqiuqiubo/andy-harness", "url": "https://github.com/andyqiuqiubo/andy-harness", "category": "Projects", "topic": "implementation-repository", "description": "Plugin-based agent framework integrating Jev structured judgments with provider configuration and session management.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3cc51449d863b9a2fd5f93c64e817c9e7a055dd9fdc964ee79be6679cf85b6e4", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "aoprisan/typesafe-ai-rust-sdk", "url": "https://github.com/aoprisan/typesafe-ai-rust-sdk", "category": "Projects", "topic": "implementation-repository", "description": "Unofficial Rust client for typed TypeSafe requests, including derived rubric structs, asynchronous calls and a blocking-client option.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5985991776a10a501962f08606c3bf81d3087bed0720758ef654cdfc94808b68", "retrieved_at": "2026-10-16T10:37:07+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Atishay8192261/card-rewards-optimizer", "url": "https://github.com/atishay8192261/card-rewards-optimizer", "category": "Projects", "topic": "implementation-repository", "description": "Self-hosted rewards calculator using Jev to interpret purchase descriptions before deterministic card-benefit calculations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7e54fe0de89d6ad1f2a812738d6cc8b3357299cf40dfff5c450e099c97ed8e39", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "austpryb/jev-hooks", "url": "https://github.com/austpryb/jev-hooks", "category": "Projects", "topic": "implementation-repository", "description": "Claude Code hooks for shell-risk screening, compaction triage and checking subagent reports through Jev.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1d502628b5e0b844360b6afb2b219ae07de9ccbaef6105e013b225b5fbb3b1db", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Barneyjm/decision-circuits", "url": "https://github.com/barneyjm/decision-circuits", "category": "Projects", "topic": "implementation-repository", "description": "Python library composes model probability answers into deterministic gates, with explicit abstention, escalation and per-item traces.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5e69c3574a1a119edb2883290e0ea1aa5aff8de9aa9a414c99579e1685c08faf", "retrieved_at": "2026-10-16T10:37:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "basi/typesafe-ai-php-client", "url": "https://github.com/basi/typesafe-ai-php-client", "category": "Projects", "topic": "implementation-repository", "description": "Community PHP client for Jev-compatible System One endpoints and configurable API gateways.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fa034a61c16b6a787dc6a3c9c3f048788c6ce8059a73401e92835cc347091cd5", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Beyond Calibration: Do a Typed-Decision Model's Probabilities Obey the Probability Axioms?", "url": "https://github.com/bro789/typed-decision-coherence", "category": "Projects", "topic": "evaluation-harness", "description": "Recorded Jev/Qwen calls and analysis scripts for probability coherence; repeated responses share source items, and post-hoc analyses are exploratory.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c6d9baee6a035052fc31c1065c52dc166cad37264b193f5b564f32d31bb88081", "retrieved_at": "2026-10-16T10:39:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "beyond-answer-confidence", "url": "https://github.com/syntheme/beyond-answer-confidence", "category": "Projects", "topic": "paper-reproducibility-code", "description": "Experiment and analysis code for Jev confidence and self-knowledge controls; datasets are fetched separately and retain mixed, sometimes noncommercial, licenses.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8d9862ab5e124dc384a0d9a2e501c1a9f635740aa0f53765263c4ae63bdf90cb", "retrieved_at": "2026-10-16T11:22:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Blakeem/logit-classifier", "url": "https://github.com/blakeem/logit-classifier", "category": "Projects", "topic": "implementation-repository", "description": "Local text-and-image classifier reads answer probabilities from logits under System One question contracts and exposes a separate abstention signal.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5985c98a6d72ba17a4d61eb25e1515fceed9e8b2f9377a03f42d0c7e17392e43", "retrieved_at": "2026-10-16T10:37:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Bongard", "url": "https://github.com/agentbull/bongard", "category": "Projects", "topic": "implementation-repository", "description": "Inference, supervised training and typed-decision evaluation code; full training data and later-stage objectives and recipes are absent.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ec8cbe0279e57ba5cee8948b859f507e942fd4e93d2bf30c284c926ca4bb92e5", "retrieved_at": "2026-10-16T10:39:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "cadence-trading/crypto-trading-lab", "url": "https://github.com/cadence-trading/crypto-trading-lab", "category": "Projects", "topic": "implementation-repository", "description": "Simulation and backtesting platform combining Jev judgments, deterministic risk vetoes and an auditable trading journal.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7591c7f80732b73779a0722097a7cb48455d92b7624f180acb0b27564ef63365", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "caglarkc/agent-browser", "url": "https://github.com/caglarkc/agent-browser", "category": "Projects", "topic": "implementation-repository", "description": "Experimental browser runtime using Jev for bounded actions over DOM observations, with published comparison artifacts.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "83bf901f81366497a97f474659e638ffbd86ae9ad44412d1ba1110bb3e6e75f2", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "call518/DeciFlow", "url": "https://github.com/call518/deciflow", "category": "Projects", "topic": "implementation-repository", "description": "Log-analysis workflow returning numeric structured judgments for batch and real-time automation through Laya or Jev.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "af1bc535e99b89695fcd193fc219a07a6718b9843952ac206f66884ee64054f1", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "capitaharlock/jev-clone", "url": "https://github.com/capitaharlock/jev-clone", "category": "Projects", "topic": "implementation-repository", "description": "Independent local decision-model implementation with dynamic options and shared-state scoring; no original Jev weights implied.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "178d68f0a387122be9752a93a7a8e7bde9b22864a6393890e56e71652866d503", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CeelNok/facet-photo", "url": "https://github.com/ceelnok/facet-photo", "category": "Projects", "topic": "implementation-repository", "description": "Photo-culling application using optional Jev judgments to apply user-defined image labels and selection criteria.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "977cdf307a110597b6df2ce1e99eef1db7c744405af01b37124f5c467d6fad2d", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "charliemartin0/testscope", "url": "https://github.com/charliemartin0/testscope", "category": "Projects", "topic": "implementation-repository", "description": "GitHub Action that asks Jev which candidate tests are relevant to a code diff.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fa7f65b13dd47bfca001265ef40d3f20d4374e310a04f46046ed1616621ad5a8", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ChenstarMr/jev-chat-windows", "url": "https://github.com/chenstarmr/jev-chat-windows", "category": "Projects", "topic": "implementation-repository", "description": "Desktop reply assistant combining screen OCR, Jev intent judgments and ranking of generated response candidates.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "768e790715748eb6c9c601f34c32296f59a07afe7518340cb3507f66baac29a6", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Chinese-Jev: Bringing System One Model to Chinese-Language Tasks", "url": "https://github.com/gulucaptain/chinese-jev", "category": "Projects", "topic": "implementation-repository", "description": "Dataset construction, fine-tuning, calibration and evaluation pipeline; the documented Chinese-Jev trained weights remain pending.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e1dedf1fa7160c9303a430b488c2702b627b0ae4aac51454a00b83677494b158", "retrieved_at": "2026-10-16T10:39:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "chkalofff/decision-qa", "url": "https://github.com/chkalofff/decision-qa", "category": "Projects", "topic": "implementation-repository", "description": "Batch question workbench for comparing local and hosted decision models through typed answers and candidate probabilities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b08f1334d3336b63bdf14c066ff27addd8848f3b000525bc65945d9af53e9b70", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "chrisDeFouRire/jevrules", "url": "https://github.com/chrisdefourire/jevrules", "category": "Projects", "topic": "implementation-repository", "description": "Tool applying shared coding rules across agents, with Jev used to check rule compliance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "bcfd719a0aca665cfd7632eb8e43c690a4454dc26c161ad733138b04219d7942", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CMaintz/classifier-bench", "url": "https://github.com/cmaintz/classifier-bench", "category": "Projects", "topic": "implementation-repository", "description": "Matched-input classifier benchmark covering accuracy, calibration, tail latency and cost across decision-model providers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f4d2d7fc7d3e504c8f3f28c4fffd735562abf24710ae7877ff0c2a4c6125e559", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "corolin/vibrato", "url": "https://github.com/corolin/vibrato", "category": "Projects", "topic": "implementation-repository", "description": "Independent emotion decision model turns message context into structured affect estimates and categorical probability outputs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "be3009dec2ed0142ef6b402f36b9e57022c4dc62889fcc32255f8ad2c4ed9409", "retrieved_at": "2026-10-16T10:37:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CR0ZER/NHL-Predictor", "url": "https://github.com/cr0zer/nhl-predictor", "category": "Projects", "topic": "implementation-repository", "description": "NHL prediction experiment combining structured statistics with LLM/Jev interpretation of news and simulated wagers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d95d0788578a1e00857a962a90c6fb8361804e2527caad4e02d20943328a1bf0", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Crosstalk", "url": "https://github.com/chaddacus/crosstalk-lab", "category": "Projects", "topic": "decision-evaluation-tool", "description": "Candidate-wording diagnostic workbench with sealed test handling; published demonstrations use a lexical mock and synthetic queries, not measured Jev behavior.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b49f69a723661bac72dab9e9d2321ea8cb5baaf5a2cad05a725498b70da1f3ab", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "CyanSalt/decision-pony", "url": "https://github.com/cyansalt/decision-pony", "category": "Projects", "topic": "implementation-repository", "description": "Compatibility layer translating between conversational model protocols and typed decision API formats.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4d6225fbb20fcc78408c6a8f396a44a44787d5de21ab61ef2469912a77c71f8b", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "cyber-security-dev-dep-mitake-com-tw/jef", "url": "https://github.com/cyber-security-dev-dep-mitake-com-tw/jef", "category": "Projects", "topic": "implementation-repository", "description": "Independent System One engine with documented temperature scaling, conformal prediction and composition of decision gates.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a5ba45c0662e8f35ad42c3a62feb472136216404b0c1effa29655a92ada41dda", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "daraujo85/jev-claude-engine", "url": "https://github.com/daraujo85/jev-claude-engine", "category": "Projects", "topic": "implementation-repository", "description": "Coding-agent integration providing Jev decision hooks, reusable skills and execution telemetry.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "57b88e3cbf783a1a600e5668af336ee160946d136d7d75cf61021604935d5c90", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "DarkisuSu/Jev-Claude-Router", "url": "https://github.com/darkisusu/jev-claude-router", "category": "Projects", "topic": "implementation-repository", "description": "Claude integration using Jev to choose relevant skills, model configuration and reasoning effort for each prompt.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e4e4e64feff593661dcd888ceffa7ddc45c5dd8b8e236ca580ec1a19d62aad66", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "davyjones7321/jev-state-engine", "url": "https://github.com/davyjones7321/jev-state-engine", "category": "Projects", "topic": "implementation-repository", "description": "LangGraph coding workflow applies deterministic validation before Jev semantic checks on changes developed in isolated Git worktrees.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "aa3c6c9d38a68c6bea9e56686cbdc24901fd1836ba987977082e190f22dc63c5", "retrieved_at": "2026-10-16T10:37:09+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "decision-models-css — social-science annotation replication package", "url": "https://github.com/hazemibrahim97/decision-models-css", "category": "Projects", "topic": "implementation-repository", "description": "Jev annotation collection, saved-call analysis and confidence-routing scripts; fresh inference needs external task data, model access and renewed API spending.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6ab7f703014db4c67112fbed8e4667d409bd3a0b3979e1257bbb08561b20977f", "retrieved_at": "2026-10-16T11:52:17.630797+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "dengyie/decidex", "url": "https://github.com/dengyie/decidex", "category": "Projects", "topic": "implementation-repository", "description": "Decision orchestration framework for Jev-style game-playing models and other bounded classifiers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e7c4af3c0517431636d5db3bf10131320ce11711b9770b6c7e9a054509467db5", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "DeRef", "url": "https://github.com/sarielma/deref_release", "category": "Projects", "topic": "paper-reproducibility-code", "description": "Pipeline builds tool-argument judge inputs, generates model-based references and scores them; model outputs and result caches are not included.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a88ad2b7483a1422ec3261e93adcab6eee1117eee2126900e3572e747da3b004", "retrieved_at": "2026-10-16T11:22:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "DigitalIQ/tabwerk", "url": "https://github.com/digitaliq/tabwerk", "category": "Projects", "topic": "implementation-repository", "description": "Browser tab-management extension using optional Jev calls for semantic questions about open tabs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0c572ab47c7291af9333acad389e48910656057a003008bd9c12a1d2cf76d4c3", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "DimeDataCloud/claude-modder", "url": "https://github.com/dimedatacloud/claude-modder", "category": "Projects", "topic": "implementation-repository", "description": "Claude Code session orchestrator using Jev to prioritize which sessions need user attention.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6e7c3864ff0dd4f761b2eb7e3fa5392dc59b31ef253f2486a7188cb492bc1193", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "docwilde/doxa", "url": "https://github.com/docwilde/doxa", "category": "Projects", "topic": "implementation-repository", "description": "Multi-agent terminal with optional Jev or SubQ decision routing and shared coordination tools.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ea501081827064142faef856292d8e49d9d2666e35c8cabe65ac8a492ab5830b", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Does a \"System One\" model behave like human System 1?", "url": "https://github.com/tynamite/evaluation-behavioral-system-one", "category": "Projects", "topic": "option-order-evaluation", "description": "Personal behavioral-evaluation archive with a preregistered option-order addendum and repeat-noise controls; it does not establish human-like psychological properties.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0cc639bdb29c1213b2e1294cda17c4af29d2d342c672710c8cb7ff48334e8723", "retrieved_at": "2026-10-16T10:37:58+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "DuvInc/jev-table-import-mapper", "url": "https://github.com/duvinc/jev-table-import-mapper", "category": "Projects", "topic": "implementation-repository", "description": "CSV-import library resolves exact column-name matches first, then asks Jev about the remaining source-to-destination column pairs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "77dd05e9917c6d16ae72dd0b16d9ef63d6ac08b29a5acb30f1ad742c3e657898", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "eaisdevelopment/jevmcp", "url": "https://github.com/eaisdevelopment/jevmcp", "category": "Projects", "topic": "implementation-repository", "description": "Coding-agent plugin using Jev to flag specification drift, failed-CI concerns and code-rule violations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "213ee53b9ef43bc3bba85bb526bb8d1148017db2ce46c0e5c29742b1ef87f1c7", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "eligrip/computer-guidance-with-jev", "url": "https://github.com/eligrip/computer-guidance-with-jev", "category": "Projects", "topic": "implementation-repository", "description": "Speech-driven Mac guidance prototype combining screenshots and repeated Jev judgments to guide interface actions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "04a91c6dbf228373ae78aae932c30256211f150d99fcad4cd71cb7580a4d6ee5", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "emo333/jev-llm-router", "url": "https://github.com/emo333/jev-llm-router", "category": "Projects", "topic": "implementation-repository", "description": "Pi extension uses Jev to select a model and reasoning budget at task entry, retaining that choice across tool continuations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7fe67557fffa71b5b5eae656d36c2597cc15dbe3c6df79b868562271eec6f3d8", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "exoticknight/dsh-system1", "url": "https://github.com/exoticknight/dsh-system1", "category": "Projects", "topic": "implementation-repository", "description": "Shared typed-decision service for DeepSeek Harness plugins, with request validation, cancellation, timeouts and pluggable model providers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2c2b3eed85d87fa63bde73282d0d4e27a9860e81f921225ec3c6aa2db1a21338", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "fanweiya/gomoku-jev", "url": "https://github.com/fanweiya/gomoku-jev", "category": "Projects", "topic": "implementation-repository", "description": "Gomoku game using Jev to choose moves against a human player.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "266332ead18369dc8539a95bc45c988f05b49dbfe24ad8e1937c7064f8961550", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Faycall1l/bouncer", "url": "https://github.com/faycall1l/bouncer", "category": "Projects", "topic": "implementation-repository", "description": "Configurable live-chat moderation application built on a Laya decision interface compatible with Jev-style questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "30500b43a564e08df8786d034c0326f2be0f6ac80f393575480d88834fe067bb", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Fox-Islam/typesafe-sdk-php", "url": "https://github.com/fox-islam/typesafe-sdk-php", "category": "Projects", "topic": "implementation-repository", "description": "Unofficial PHP client offers fluent typed-question construction and switching between TypeSafe and OpenRouter endpoints.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "709a1491bfc04504fc3da1780f386855061831c336ac02f65d8aee950ca8c849", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "FrantisekJanos/PickBench", "url": "https://github.com/frantisekjanos/pickbench", "category": "Projects", "topic": "implementation-repository", "description": "Decision-model benchmark combines interactive games with measurements of latency, confidence calibration, prompt variation and injection resistance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a3d26e87fe06d0f10a46cb9ef3cc1a48eb4e950e8c4f9421c21d8ffc10475394", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "frederico-kluser/dwarf-fortress-agent-skill", "url": "https://github.com/frederico-kluser/dwarf-fortress-agent-skill", "category": "Projects", "topic": "implementation-repository", "description": "Dwarf Fortress agent combining Jev decision batches, a language-model narrator and campaign memory.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7fc934720066c07f18316349f0ff1360a7fcc0cb66c0909233c3e94e2a3f14d1", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "fstandhartinger/auto-model-router", "url": "https://github.com/fstandhartinger/auto-model-router", "category": "Projects", "topic": "implementation-repository", "description": "LLM router combining capability, cache and quota constraints with Laya, hosted Jev or heuristic request classification.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9386ab242e35cb39036e3228061ba6503a4229c6c637398023b20f5a39a35fdd", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "GabrielCoelhoCruz/jev-scanr", "url": "https://github.com/gabrielcoelhocruz/jev-scanr", "category": "Projects", "topic": "implementation-repository", "description": "Code-refactoring assistant using Jev judgments to prioritize possible duplication and separable responsibilities in functions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "20706b2539fc0b5e54f9b760fd2dff7f78c7bbe52062570667fd3aefd3e0da61", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "GabrielCoelhoCruz/jev-skill-router", "url": "https://github.com/gabrielcoelhocruz/jev-skill-router", "category": "Projects", "topic": "implementation-repository", "description": "Auditable skill-routing implementation comparing Jev decisions with an offline lexical baseline.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5f8d36417983f194372bb5f1488289526d4373ee4d2873fadd680f95ed09d002", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Giamme/apple-silicon-systemone-benchmark", "url": "https://github.com/giamme/apple-silicon-systemone-benchmark", "category": "Projects", "topic": "implementation-repository", "description": "Apple Silicon harness compares local decision models and hosted services on shared labeled cases, exporting dashboards and request-level records.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b62e97c372295a8450a983196de50e6b56a9818e1ea22d18269bb12e23fb6bad", "retrieved_at": "2026-10-16T10:37:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "grafuls/huncho", "url": "https://github.com/grafuls/huncho", "category": "Projects", "topic": "implementation-repository", "description": "Rust inference server hosts local decision-model families behind a TypeSafe-compatible HTTP endpoint and supports multiple execution backends.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8158a36169367dd53bce6921ead43f292b227d3d3679f8568ac1b38e85d3f687", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "gsaini/system-one-models-getting-started", "url": "https://github.com/gsaini/system-one-models-getting-started", "category": "Projects", "topic": "implementation-repository", "description": "Runnable experiments examining System One launch claims through shared typed interfaces, calibration tests and task examples.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4976c081008bf119d554542d9d16f7956c478615bd2ec612efb9a6119ab7db23", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "guow7394/ecom-cs-agent", "url": "https://github.com/guow7394/ecom-cs-agent", "category": "Projects", "topic": "implementation-repository", "description": "Customer-support agent using Jev intent and escalation judgments before selecting automation or human review.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5738fc89920424f0b6f336d1bc932e93f18f329fbf1ac51f35b8198802f95577", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "hajdu-patrik/trirouter", "url": "https://github.com/hajdu-patrik/trirouter", "category": "Projects", "topic": "implementation-repository", "description": "Multi-model coding workspace with prompt routing and optional Jev or local classification.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d31dfcd45f1377f9ec51f15cf4086946e4059f56e162a93cada3457642665182", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "hakkisagdic/neo-vs-morpheus", "url": "https://github.com/hakkisagdic/neo-vs-morpheus", "category": "Projects", "topic": "implementation-repository", "description": "Game-duel agent using local Laya or hosted Jev decisions, with a live monitor for the resulting behavior.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1d4bc320e4a59c591e510b8f70f32b7403cbff9126be10fe5e8485a13a8ee85f", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "hellolitebox/litebox-inbox-jev", "url": "https://github.com/hellolitebox/litebox-inbox-jev", "category": "Projects", "topic": "implementation-repository", "description": "Email-triage application using Jev classifications to drive configurable labeling and inbox actions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9b799146e3378d07ccd8cd8709e22e92e73e2da5999e43c1a98a0670c3217929", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Hetgandhi25/jev-decision-studio", "url": "https://github.com/hetgandhi25/jev-decision-studio", "category": "Projects", "topic": "implementation-repository", "description": "Incident-triage application combines local Laya classification with an LLM fallback to produce category, urgency and severity fields.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e24b4395466371ec31ae227353c22af441c45e440922d6d1e1f353ed673bddcd", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "higress-group/HiRoute", "url": "https://github.com/higress-group/hiroute", "category": "Projects", "topic": "implementation-repository", "description": "Local agent-coordination engine supporting configurable decision-model providers, including Jev-compatible endpoints.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "746427780065f67a8b0ad1e2db3882bf4fb139b5392caaea1a4c62f507900431", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "HillPhelmuth/JevDotNet", "url": "https://github.com/hillphelmuth/jevdotnet", "category": "Projects", "topic": "implementation-repository", "description": "Community C# SDK for System One decision APIs, with configurable providers and model versions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "470950db0f5e3085633e3e9d397f944a00080ff0cc165497c52da726dbf9e509", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Hyper-AI-Lab/openclaw-jev", "url": "https://github.com/hyper-ai-lab/openclaw-jev", "category": "Projects", "topic": "implementation-repository", "description": "OpenClaw control plane using Jev intake routing, workflow orchestration and a response-evaluation gate.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "dcc5cf09b3ff2ca658c27d797e746ce2d9a88d2a6bab160d4ef8e27950ca8d5e", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "iamngoni/veyra", "url": "https://github.com/iamngoni/veyra", "category": "Projects", "topic": "implementation-repository", "description": "Trading-service research implementation placing typed Jev judgments behind deterministic order-risk checks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f2606840af7a4b9e0edeb902b6bbf281838400576e363db7bfce2d818eee6aef", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "igrady2/cant-bleep-this", "url": "https://github.com/igrady2/cant-bleep-this", "category": "Projects", "topic": "implementation-repository", "description": "Class-forum chat demonstration that applies Jev moderation before adding messages to the shared log.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c5efb0014870475a334445a160743ccb1cc5f914427523a36f20cddf7a906746", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "IoTone/jevracket", "url": "https://github.com/iotone/jevracket", "category": "Projects", "topic": "implementation-repository", "description": "Racket-language client for the hosted TypeSafe typed-decision API.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "17b41b033a6af576c03117960e9605b11385787787c0a10a152d19f2f808033b", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "irenetong904/spotify-review-pipeline", "url": "https://github.com/irenetong904/spotify-review-pipeline", "category": "Projects", "topic": "implementation-repository", "description": "Review-analysis pipeline separating fixed-choice Jev classification from language-model synthesis and independent checks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "c163dfec6a90d5d9d6899850887cd83ad2bebc255a738b3b679d89d944cf7a62", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "itsamejoshab/jevlab", "url": "https://github.com/itsamejoshab/jevlab", "category": "Projects", "topic": "implementation-repository", "description": "Offline phrase-search laboratory for the Trick Jev classifier, with scored candidate histories and adaptive search allocation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6758eedfb1b8c21acef35a0ff2ccfd7f5d96aafcbc6d51688c12e67a024133e0", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "izam-mohammed/decisionsmith", "url": "https://github.com/izam-mohammed/decisionsmith", "category": "Projects", "topic": "implementation-repository", "description": "Decision-model toolkit for teacher-guided data preparation and fine-tuning workflows involving Jev and Laya.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2f8c7dcf2591366c38f83e58e92acab0f0baf2b4348fb45a75e1d453e7eb8908", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "j62268781-alt/dsh-judgment", "url": "https://github.com/j62268781-alt/dsh-judgment", "category": "Projects", "topic": "implementation-repository", "description": "Agent plugin exposing batched typed judgment tools with response validation, audit records and budget accounting.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "23820f13a139db63727ccfc8b865bd13bee5fcbb6272738294d95307f665db03", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jackthony/IA-local", "url": "https://github.com/jackthony/ia-local", "category": "Projects", "topic": "implementation-repository", "description": "Local-agent laboratory evaluating decision routing and refund policies with labeled cases, calibration and human escalation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "14f84ae77bae175c3a7037f9ed806aaff7470baeb1440d8f96b6d3a071b2fa65", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jasonduncan/semantic-operators", "url": "https://github.com/jasonduncan/semantic-operators", "category": "Projects", "topic": "implementation-repository", "description": "Provider-neutral Python, CLI and MCP abstractions for typed questions, answer probabilities and uncertainty handling.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "43f0e21408cde59bfbb44be3a7139a4addcc666ceb14cf224907199787450265", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jerichosiahaya/naluri", "url": "https://github.com/jerichosiahaya/naluri", "category": "Projects", "topic": "implementation-repository", "description": "XLM-RoBERTa-based multilingual decision model exposes yes/no, choice and ordinal questions, with fine-tuning and evaluation guidance.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e9c379f2bd7120d0d9b32512430b9c5f9c930910b40a3cb30ef1d13807354fd7", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev-Benchmark — contract inference records and analysis", "url": "https://github.com/zf-utokyo/jev-benchmark", "category": "Projects", "topic": "implementation-repository", "description": "ContractNLI scoring and sanitized prediction records for the initial comparison; raw API bodies and the full revised-paper follow-up are outside the documented release scope.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b6941cbdf02a8922e12a58bfc853b1be2c8f2182148e28d7290c275961770f64", "retrieved_at": "2026-10-16T11:57:24.072123+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Jev-Sinhala Capability, Primitive Consistency & Benchmark Census Probe", "url": "https://github.com/azri-muhsin/jev-sinhala-review", "category": "Projects", "topic": "multilingual-robustness-evaluation", "description": "Sinhala and Singlish evaluation using established benchmark labels, with script changes, question-language controls, option permutations and cross-primitive consistency tests.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "30558333a111edb418272359343e250d82ef45323ec5a4505fd0e0c96020d214", "retrieved_at": "2026-10-16T10:37:58+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JEV-Star", "url": "https://github.com/sc2musa/jev_star", "category": "Projects", "topic": "implementation-repository", "description": "StarCraft II controllers and selected comparison artifacts; full original logs are excluded, version changes confound planning-effect claims, and live runs need game/model services.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d2d620b5142438143329c8cd524bed46b7d0f59707ffaa200c4a2a9a1d074e54", "retrieved_at": "2026-10-16T11:04:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jev-xray", "url": "https://github.com/aishwary-dongre/jev-xray", "category": "Projects", "topic": "decision-evaluation-tool", "description": "Input-ablation tool measuring changes in typed decision probabilities; useful for testing distracting context without treating attribution as a causal explanation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "18a9f5a7fb939dc63c23153ace40d5f4486f7d490180b31be3b0d9364fc4e822", "retrieved_at": "2026-10-16T10:37:58+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JevAdvBench — attack generation and offline analysis", "url": "https://github.com/jevadvbench/jevadvbench", "category": "Projects", "topic": "evaluation-harness", "description": "Attack-generation and offline analysis code for JevAdvBench separates recorded-response replay from fresh hosted inference; labels are mostly model-derived, and new API runs need not repeat saved answers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9dd2b331797756acd01229ca3aaed939a91518912410f49e6bdb7cb288548469", "retrieved_at": "2026-10-16T12:54:12.571910+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jhochenbaum/rlm-jev", "url": "https://github.com/jhochenbaum/rlm-jev", "category": "Projects", "topic": "implementation-repository", "description": "Recursive-language-model integration exposing Jev classification and location judgments as bounded sub-calls.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5920f9f9c91113e0d2fcba522a93855b867f4e2e78202653b3174daec68c4070", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jimador/bayes-of-our-lives", "url": "https://github.com/jimador/bayes-of-our-lives", "category": "Projects", "topic": "implementation-repository", "description": "Embabel example comparing interchangeable Jev, OpenAI and deterministic local decision services.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2124acbfe9ca111f36f6ab0720dbc8b797d4190caa74bc83b16a00cbba1fd40a", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "jonny5isalive5/jev-choice-probability-consistency", "url": "https://github.com/jonny5isalive5/jev-choice-probability-consistency", "category": "Projects", "topic": "implementation-repository", "description": "Frozen exploratory Choice-probability experiments preserving complete results, failures and reproducible diagnostics.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "125e5bde84a65467f2f3c5a83977bec0a32745c74632ad65b18f748d124a79e1", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "juspay/neurolink", "url": "https://github.com/juspay/neurolink", "category": "Projects", "topic": "implementation-repository", "description": "Multi-provider AI library with a separate typed decision interface for TypeSafe Jev alongside text generation.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d4db0a2f1b92a987b2bea7874ac71df0c265bb94409205520d242062802abfeb", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "justintout/systemone", "url": "https://github.com/justintout/systemone", "category": "Projects", "topic": "implementation-repository", "description": "Unofficial Go module for Jev and compatible System One models, with environment-based client configuration.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3a5ed3a7525f28ad50d5e47d7859dbb38117a0df031155ea0040779828473b33", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "JYeswak/jev_playground", "url": "https://github.com/jyeswak/jev_playground", "category": "Projects", "topic": "implementation-repository", "description": "Coding-agent classifier research workspace with labeled evaluations, decision logs and explicit promotion criteria.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2d6522c531f071db0d5ff46ce13cbb4085036fd1a986cf1e358f0ff0cae84717", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "khs0927/jev-browser-control-plane", "url": "https://github.com/khs0927/jev-browser-control-plane", "category": "Projects", "topic": "implementation-repository", "description": "Browser-control prototype integrating Jev decisions with existing browser harnesses and limited coordination code.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "acc047009aeb08a092be3812125f8ba044e3ae2d61a8fe24b6e0b5e46da4264f", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "KITE — population experiment code and aggregate evidence", "url": "https://github.com/hengyuli-ozaki-lab/kite_population_simulator", "category": "Projects", "topic": "implementation-repository", "description": "Population-kernel simulation and aggregate evaluation records; survey microdata and Jev responses are withheld, and hosted-model reruns need not reproduce saved outputs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e008e4aeadc9bb2e941a1e67682b89583d8963a29704a8c4a3217d4ce54cb7ad", "retrieved_at": "2026-10-16T11:57:23.819716+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "kitfunso/hippo-memory", "url": "https://github.com/kitfunso/hippo-memory", "category": "Projects", "topic": "implementation-repository", "description": "Coding-agent memory system offering optional Jev reranking of retrieved context.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4e3d0a83d605a41588441fcbc0f345e55be31c839ba81651a5ba1859748433e6", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "kkdev92/jev-dotnet", "url": "https://github.com/kkdev92/jev-dotnet", "category": "Projects", "topic": "implementation-repository", "description": "Unofficial .NET client with typed decision plans, contract-checked answers and Native AOT support.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "089dfb3fa2786145e8c5ebf70cb9a0756bc534ed3fc204c34ed9272e62829d30", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "klnstprx/typesafe-sdk-go", "url": "https://github.com/klnstprx/typesafe-sdk-go", "category": "Projects", "topic": "implementation-repository", "description": "Community Go library for TypeSafe questions and typed responses, with configurable retries, custom response types and logging.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "80a7e79b584b9a0375c51ad2ff4034752a125590ce4f7408b328cf7884587489", "retrieved_at": "2026-10-16T10:37:07+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "kodr-pro/AiWrangler", "url": "https://github.com/kodr-pro/aiwrangler", "category": "Projects", "topic": "implementation-repository", "description": "Coding guardrail tool using Jev judgments, with an explicit option to disable external inference.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5ba25733f03c55ccc063739f28600e77c9b0531d7014ffd4b537c6249c33e510", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Laya benchmark and PoC", "url": "https://github.com/banhun28/laya-benchmark-poc", "category": "Projects", "topic": "multilingual-robustness-evaluation", "description": "Korean decision-model study with matched Jev/Laya comparisons, fine-tuning controls and option-order probes; published replication limitations remain explicit.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "dbaeb1eab4bb7c638b05c82ff22f81faec6047e3d6a49d767cc1ac7e40152a16", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Laya zero-shot agent-step guard evaluation", "url": "https://github.com/zephyr4772/laya-zeroshot-agent-guard-eval", "category": "Projects", "topic": "decision-robustness-benchmark", "description": "Small handwritten zero-shot diagnostic of agent-step decisions; reported failures are specific to the tested prompts and model configuration.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2a3abd27037c89beec0241e5d4ed8ab747e64b8528b8030f39ccd95a72b4de5a", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "laya-rlcd-training", "url": "https://huggingface.co/InfinimindCreations/laya-rlcd-training", "category": "Projects", "topic": "implementation-repository", "description": "Public implementation of a proposed Laya RLCD training loop, with corpus experiments and optimization details.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "cbf1c15b236b4991a9a0e50c06f4c4ffa21fb3751a727c6c62347f9bd0431432", "retrieved_at": "2026-10-16T10:05:23+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "lesnolie/magpie-free-gateways", "url": "https://github.com/lesnolie/magpie-free-gateways", "category": "Projects", "topic": "implementation-repository", "description": "Gateway plugin exposing Jev as a native classifier within Magpie's routing interface.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "501d426b39df6a1e1e08cef95bb3efa48ce84ba993ae2d1c7f2045808c2694b1", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "LightJev", "url": "https://github.com/rongxinzy/lightjev", "category": "Projects", "topic": "implementation-repository", "description": "Finite-candidate scoring, CE/Brier training and temperature fitting; published results cover synthetic tasks and one seed, not proprietary RLCD reproduction.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "5b984a917c8ac676e2f51ded74eea739d6756e12e1016a7f0d97ae6e675b6fab", "retrieved_at": "2026-10-16T10:39:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "LLMs for CSS — upstream task data and evaluation harness", "url": "https://github.com/salt-nlp/llms_for_css", "category": "Projects", "topic": "benchmark-source", "description": "Task data, prompt mappings and evaluation utilities explicitly required by the Jev social-science replication package; upstream benchmark material, not separate Jev results, and no top-level license was found.", "relation": "adjacent", "verification_level": "content_screened", "content_sha256": "698568f67a5ff008970e4eb75e8fc3d4fe021b57bb611056afd1076063ea75c8", "retrieved_at": "2026-10-16T11:52:17.773226+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "lostcol0ny/lwc-hoa-qa", "url": "https://github.com/lostcol0ny/lwc-hoa-qa", "category": "Projects", "topic": "implementation-repository", "description": "Document-question-answering system combining Jev retrieval decisions with generated answers and citation verification.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9cb27a3172254a1dea63a88679634cef1aec24a34ea79febd74993a76b07232e", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "lymagics/jev-gate", "url": "https://github.com/lymagics/jev-gate", "category": "Projects", "topic": "implementation-repository", "description": "CI action translating code-quality questions into Jev probability thresholds that can fail a pipeline.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "49b1cd70eb0b100d239e414d45e35460b7af33f9bc530dccb0bd330032c60ef9", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "lytang63/OmniMed-Jev", "url": "https://github.com/lytang63/omnimed-jev", "category": "Projects", "topic": "implementation-repository", "description": "OmniMed-Jev research code studying calibrated multimodal medical decision interfaces; clinical validity is not established here.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "11b53aa30799c5ef0049f3c96dfb1395bd54e32be28fb0d43ddfea4b18bd7081", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "maloosatyam-google/apigee-ai-gateway-reference", "url": "https://github.com/maloosatyam-google/apigee-ai-gateway-reference", "category": "Projects", "topic": "implementation-repository", "description": "Apigee gateway demonstration classifies prompts with Jev, then maps categories to model providers through API-product policies and quotas.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6d60bb5ee57616d44d72fe31cd94d21072f7f93c4c4f61dd893dc0082e3056f0", "retrieved_at": "2026-10-16T10:37:10+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "maneeshmenongit/model-selector", "url": "https://github.com/maneeshmenongit/model-selector", "category": "Projects", "topic": "implementation-repository", "description": "Model-selection service using Jev to choose a capability tier and LiteLLM to select a model within it.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1b7fe1e354ba6be69a2d924cdd2d9b537c1f1b320251e1df85235695b9b7af54", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "MarcelRoozekrans/Minos.NET", "url": "https://github.com/marcelroozekrans/minos.net", "category": "Projects", "topic": "implementation-repository", "description": "Source-generated .NET decision abstractions spanning hosted providers and local System One-compatible servers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "aaacc232b063dd2a30e4835ca86f9cda9212d2a934a6449d74d0a33056cedb2d", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "max1874/open-computer-use", "url": "https://github.com/max1874/open-computer-use", "category": "Projects", "topic": "implementation-repository", "description": "Mac computer-use port mapping accessibility-tree elements into bounded actions for Jev or conversational model backends.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "26769c5fb95e2a5e9de62a689afa58a4ae880909aaebca789c97c9870502f23e", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "mcontre4/jev-trader", "url": "https://github.com/mcontre4/jev-trader", "category": "Projects", "topic": "implementation-repository", "description": "Experimental Jev trading terminal with preview mode, risk checks and execution records.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fe38bcc7f5c030941141a4e7bf5de9e5816dcd233b0f65a51cfb052a26cad139", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "mgd34msu/goodvibes-jev", "url": "https://github.com/mgd34msu/goodvibes-jev", "category": "Projects", "topic": "implementation-repository", "description": "In-progress agent-platform rebuild with an implemented shared judgment transport; broader autonomous-decision goals remain under development.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2746c5bf0ac463112f70d54689fca277a13bb4c7bac432889e7dd4c25f85238d", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "milhous/turnhelm", "url": "https://github.com/milhous/turnhelm", "category": "Projects", "topic": "implementation-repository", "description": "Local-first CLI using Jev or Laya to select a coding model and reasoning effort.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ca743d57d57d8898607831640cba5f9797cc6b4b77e553d8769bbcefbc4a98ac", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Mindbreaker81/system-one-bench-public", "url": "https://github.com/mindbreaker81/system-one-bench-public", "category": "Projects", "topic": "implementation-repository", "description": "Public benchmark release comparing typed models on respiratory-domain triage, literature classification and manipulated messages.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fd53f940bd5d0f065c7729bf0119f651ca8c0b31919dac1b50a815bb0170e970", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "MohibShaikh/jev-skillbench", "url": "https://github.com/mohibshaikh/jev-skillbench", "category": "Projects", "topic": "implementation-repository", "description": "Malicious agent-skill detection benchmark using Jev with a verification-and-escalation cascade.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "717d5913f15abea6df45925d94176ac6a31717a1309f74445e11e810169e22d9", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "monotykamary/reflex", "url": "https://github.com/monotykamary/reflex", "category": "Projects", "topic": "implementation-repository", "description": "Multi-agent desktop control interface using a System One router to direct requests among coding agents.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "684eab4be787d0171c6e6cb1d2f08ba17310bd49f2470a403f3eeb2d024fa033", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "mrz1836/go-discern", "url": "https://github.com/mrz1836/go-discern", "category": "Projects", "topic": "implementation-repository", "description": "Typed Go decision layer covering TypeSafe Jev and Cloudflare Clef APIs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7cd84c6eb3b554a35f684a9da60f93d5b93abfba3b473ea73ca6d0efbb4788bd", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "munod/tachyone", "url": "https://github.com/munod/tachyone", "category": "Projects", "topic": "implementation-repository", "description": "Independent multilingual local decision engine serving the Jev-compatible typed-question protocol.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "14b20d7ddb087419d0b99549e0079552b87ad6fa1954a675c09a721fa9c4278b", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "nanagb/jev-frame-engineering", "url": "https://github.com/nanagb/jev-frame-engineering", "category": "Projects", "topic": "implementation-repository", "description": "Coding-agent plugin for designing, testing and refining structured decision frames for Jev integrations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a3963ed6ba5f836c58f70541f5524a80a601150c55e052a208d87cf62524e288", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "nanodiff-350m-typed-decisions", "url": "https://huggingface.co/pngwn/nanodiff-350m-typed-decisions", "category": "Projects", "topic": "implementation-repository", "description": "Masked-diffusion typed-decision experiment code with a synthetic generator and calibrated-scoring study.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1de49c73e1dd70a7fb45b498c534f23653afa589117a855c75b4d1a561374696", "retrieved_at": "2026-10-16T10:05:29+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "nedcut/gm-bench", "url": "https://github.com/nedcut/gm-bench", "category": "Projects", "topic": "implementation-repository", "description": "Multi-season sports-management benchmark with a separate Jev question-scaffold lane, documented apart from its preregistered chat-model comparison.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0a034c829b991948acb4170b8e735f036bb7977af15756efcd1668832e30b4fd", "retrieved_at": "2026-10-16T10:37:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "neostryder/augur", "url": "https://github.com/neostryder/augur", "category": "Projects", "topic": "implementation-repository", "description": "Provider-usage application with optional Jev rubric grading and a local gateway for typed decisions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "994b1bb9855a61cc0398e8e96a49560931e09da2d58d4dacdac684632745e7f9", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "NMTZ-z/system-one-benchmark-lab", "url": "https://github.com/nmtz-z/system-one-benchmark-lab", "category": "Projects", "topic": "implementation-repository", "description": "Local agent-control runtime combines Laya decisions, an optional Apple Neural Engine backend, policy gates and platform-specific adapters.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e2b48fec25acf98de7a12784221178b716461c5a622b399d2ea49d92c89c4cd6", "retrieved_at": "2026-10-16T10:37:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "notque/vexjoy-agent", "url": "https://github.com/notque/vexjoy-agent", "category": "Projects", "topic": "implementation-repository", "description": "Agent dispatcher using Jev to classify requests and check that selected workflows preserve the intended outcome.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3e8684fca860b3b01ca44aa5cbb7015e163a67dec3ebe91677e1041253bb5d58", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "nvr0x5/claude-deck", "url": "https://github.com/nvr0x5/claude-deck", "category": "Projects", "topic": "implementation-repository", "description": "Claude Code cockpit using Jev to select subagent models and per-prompt reasoning effort.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d17a91666f9a9ec3ae2e7afa242183f8c707c3a25dda87de9166364bd02da3ae", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "obekt/obekt-terminal", "url": "https://github.com/obekt/obekt-terminal", "category": "Projects", "topic": "implementation-repository", "description": "Experimental autonomous-trading platform using typed Jev judgments behind code-defined risk controls and a monitoring terminal.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a4b828f0409516def99995e6abf493806652c452cb006a5a220a1905b700c14a", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "One Word Opens the Gate", "url": "https://github.com/arminazizi98/option-channel-attack", "category": "Projects", "topic": "paper-reproducibility-code", "description": "GuardBench generator and cached open-weight guardrail evaluations; not a hosted-Jev exploit demonstration, and synthetic policy labels do not establish real-world attack rates.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4ffb4a18b98a5c90c8b7c34521b636fce7f2a1ecfacfeac86c3cee4b76b92805", "retrieved_at": "2026-10-16T11:22:28+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "OpenByteInc/QuantDinger", "url": "https://github.com/openbyteinc/quantdinger", "category": "Projects", "topic": "implementation-repository", "description": "Trading research platform integrating Jev judgments with strategy context, exposure and budget checks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2537041ca9038b022b25318fe7c2002738da6f44bcb7be8e08f4be00f8095bf9", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Ordinal-scale utilization — anonymous review artifact", "url": "https://anonymous.4open.science/r/jev_ordinal_scale_bia", "category": "Projects", "topic": "implementation-repository", "description": "Anonymous ordinal-scale audit code, frozen indexes and documented prediction releases; original text and post-trained checkpoints are omitted, and unavailable historical data revisions can block exact reconstruction.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d56870df6300d1ea9c1a303712a7c12dec59e140880e057ef00e9ff1c46b37bc", "retrieved_at": "2026-10-16T12:22:44.534032+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "OutThisLife/jev-tutorial", "url": "https://github.com/outthislife/jev-tutorial", "category": "Projects", "topic": "implementation-repository", "description": "Jev tutorial and examples covering workflow gates, automation integrations and real-data exercises.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "529f23c2781b0cc090bbc824a373d516459a841819777aa21170bac833d42bd2", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "PabloB07/JevMC", "url": "https://github.com/pablob07/jevmc", "category": "Projects", "topic": "implementation-repository", "description": "Unofficial Java System One client with typed wrappers and a game-server chat-moderation integration.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "992bcef1db7b8458304ed576de62a18bbe6d843ff1d515a8a8c69298dc07cb83", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "pachelbel23/claude-code-jev", "url": "https://github.com/pachelbel23/claude-code-jev", "category": "Projects", "topic": "implementation-repository", "description": "Claude Code plugin providing Jev task, skill and shell-risk hooks, plus vision-question support.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0cd8764950620357ca8544dafbb049bc40d7396e2b51eba988ef41b9a9c32725", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "PanikParty/System1Doom", "url": "https://github.com/panikparty/system1doom", "category": "Projects", "topic": "implementation-repository", "description": "ViZDoom agent combining typed decisions from Jev or Laya with a deterministic game controller.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b241f779fa7f7880c76eed1e79989dc6077f08419225a1e85456058f8ddfb4b5", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Parcha-ai/agentrun", "url": "https://github.com/parcha-ai/agentrun", "category": "Projects", "topic": "implementation-repository", "description": "Workflow language connects Jev judgments, application tools and delegated agents through explicit adapters, with offline fixtures and live-provider integration.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b6317b10d64a855efbfe28caf6c79932cfd078497bbffaaac6f63a57050c7003", "retrieved_at": "2026-10-16T10:37:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "patkusch/assay", "url": "https://github.com/patkusch/assay", "category": "Projects", "topic": "implementation-repository", "description": "Independent local System One implementation for bounded decision questions and confidence-aware outputs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6416193fcc41085fb50ccd63617e17733aff6cc3d610dd946177710c9caad1d5", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ppweni/typesafe_rs", "url": "https://github.com/ppweni/typesafe_rs", "category": "Projects", "topic": "implementation-repository", "description": "Asynchronous Rust client supports TypeSafe text, image and Box questions, with typed response handling and configurable transport behavior.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "56bffb38356f9d410e61f71ee027fe2b60d21f15548b96c51d2074b55edf79a3", "retrieved_at": "2026-10-16T10:37:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "pyck-ai/jev-cli", "url": "https://github.com/pyck-ai/jev-cli", "category": "Projects", "topic": "implementation-repository", "description": "Command-line and MCP tools exposing hosted Jev typed judgments through OpenRouter.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "bd2017d019d6f7ccb205de2013425adf42b3e7b0179b72bd19cec9c813351ca8", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "raj8525/universal-jev", "url": "https://github.com/raj8525/universal-jev", "category": "Projects", "topic": "implementation-repository", "description": "Coding-agent runtime combining bounded Jev browser actions with a higher-level browser-agent fallback.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "63aed3a3aa3c56fe750f77da742bab80bc315203d3c513e74f7b377ad27e9c83", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "raullenchai/Rapid-MLX", "url": "https://github.com/raullenchai/rapid-mlx", "category": "Projects", "topic": "implementation-repository", "description": "Apple Silicon inference server includes MLX-based System One serving for Laya and other decision-model backbones.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ae66c7a35722410c670b4274c8074b353f5163b0b1edf77f5b115fc015f1a18d", "retrieved_at": "2026-10-16T10:37:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "rayopavri/jev-setup", "url": "https://github.com/rayopavri/jev-setup", "category": "Projects", "topic": "implementation-repository", "description": "Claude prompt hook routing work between hosted and open-model lanes using typed Jev judgments.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2ece156923de785155b3b0e7c7f79ff13a5679a3ed2a8b0121b9591417e27cf1", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "receptron/kenteiscript", "url": "https://github.com/receptron/kenteiscript", "category": "Projects", "topic": "implementation-repository", "description": "Quiz DSL and runtime using Jev to grade free-form learner answers against question-specific criteria.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "85b2f127731756702bf9ccaa9a0a90fa0eccfd4ec12c7fa49220fae0c3462902", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ricky94511/jev", "url": "https://github.com/ricky94511/jev", "category": "Projects", "topic": "implementation-repository", "description": "Claude Code plugin for Jev-based model-tier selection and typed classification, scoring and yes/no questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "122dc83c9e44344dddd2f6a4e93e2cd84ffa79b5966a86a89ca6e7fb2c6c8162", "retrieved_at": "2026-10-16T10:05:40+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "robertanton81/claude-subagent-router", "url": "https://github.com/robertanton81/claude-subagent-router", "category": "Projects", "topic": "implementation-repository", "description": "Claude subagent plugin with optional Jev routing of task briefs and a separate Codex bridge.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1158d9bcfbe82eb6a0452855eed140eea75ea7a9f6949fd70e169f8af68ae679", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ruvnet/RuVector", "url": "https://github.com/ruvnet/ruvector", "category": "Projects", "topic": "implementation-repository", "description": "Local decision and memory framework with labeled evaluation, abstention information and a Jev-compatible serving option.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7ef50964ae9b77d8fdcac9f4ce49faa7eeb7c98c3b972137cadd2c7765787c37", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SafeTicket", "url": "https://github.com/cyber-traveller/safeticket", "category": "Projects", "topic": "robustness-policy-harness", "description": "Support-routing policy harness testing option permutations, ambiguous tickets and injection review gates; fixtures are authored synthetic cases, not independent customer evidence.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "da43ec756e3eb625c408f0a22ae026a4cefac8a7706272112adb6a3592578c6d", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "sakhadib/literal-lock-in", "url": "https://github.com/sakhadib/literal-lock-in", "category": "Projects", "topic": "multilingual-robustness-evaluation", "description": "Bengali idiom study controlling literal distractors, option labels and question language while reusing expert figurative-meaning and literal-gloss annotations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d3f0869ae9db200d746bec6ae6d9cfde906fbc16511dca0ad3544f0623d91061", "retrieved_at": "2026-10-16T10:37:58+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "sathiskumarjothi0-oss/bev-decisions-pypi", "url": "https://github.com/sathiskumarjothi0-oss/bev-decisions-pypi", "category": "Projects", "topic": "implementation-repository", "description": "Python package serves Beveridge model variants locally through yes/no, categorical-choice and ordinal-score interfaces.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8a14cc11de80c62b8b98c3cc26dc14412e6b10287ae9da6dfaff0ad19f1cbb45", "retrieved_at": "2026-10-16T10:37:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "schalkneethling/jev-3000", "url": "https://github.com/schalkneethling/jev-3000", "category": "Projects", "topic": "implementation-repository", "description": "Interactive workbench comparing typed requests across Jev and other decision providers, including image inputs.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8d92da6a9b084f0ddc34e1f5820bcdaabd85a295b2d0f4dc8fab467bf5376b1d", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Scientific Decision Evaluation — author code and prepared-case artifacts", "url": "https://github.com/bydeng01/scientific-decision-eval", "category": "Projects", "topic": "implementation-repository", "description": "Scientific Choice runners and saved-result checks; current expanded development cases differ from the paper snapshot, and repeated prepared cases do not establish generalization.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8f1c374132448584112508810e7ee024975bc6d4f581bf79116f0612bd786433", "retrieved_at": "2026-10-16T11:57:23.843862+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "sdogantekin/jev-orchestrator", "url": "https://github.com/sdogantekin/jev-orchestrator", "category": "Projects", "topic": "implementation-repository", "description": "Simulation-only payment-routing laboratory comparing Jev decisions with deterministic rules; no real payment execution.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "f49ebc1a2e578dc64dd986b18e53c0ec0f07cabe2e9c9cddcbfcfb2443a35c86", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Selective Reasoning for Typed Decision Models", "url": "https://github.com/binzheng-joey/selective-reasoning", "category": "Projects", "topic": "implementation-repository", "description": "Candidate-readout policies and saved-measurement recomputation; fresh inference still needs external weights, benchmark data and feature caches.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "4baa6da446c3b0eca4efb07a625810f30f61bc614a134c3f1d6c3c1d463e207e", "retrieved_at": "2026-10-16T11:04:25+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "seominpapa/siteeye", "url": "https://github.com/seominpapa/siteeye", "category": "Projects", "topic": "implementation-repository", "description": "Construction-site research prototype combining vision detections with Jev near-miss judgments; safety effectiveness is not established here.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "070d96aaea29b9420369f13a9e28f15fe6c1168730cacf948967f8bf3a3647b9", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SergeyBazhmin/krasivo-jev", "url": "https://github.com/sergeybazhmin/krasivo-jev", "category": "Projects", "topic": "implementation-repository", "description": "Dataset conversion, training and evaluation tools for single-pass multiple-choice decision models.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "caa27e2fb847ced11b8ffca251f63e2e632c54883e9440e2aa372c1bcca3772d", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "sgyaqing/verdict", "url": "https://github.com/sgyaqing/verdict", "category": "Projects", "topic": "implementation-repository", "description": "Serving engine for models using the System One typed-decision protocol.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "02875cb68e5abda3a9d06b205aefa7dd6f30f3e35c8b020d40a63f54858ebf23", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "shivanathd/gavel", "url": "https://github.com/shivanathd/gavel", "category": "Projects", "topic": "implementation-repository", "description": "Salesforce integration applying versioned Jev judgments through governed decision workflows.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3565ac062e191d99b2032c89921cc99292603531de35160b53dbec25ea4b1adf", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "shivanathd/jev-playground", "url": "https://github.com/shivanathd/jev-playground", "category": "Projects", "topic": "implementation-repository", "description": "Browser playground that exposes Jev probabilities, confidence thresholds and request/response JSON across typed-question workflow examples.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3d6f752acd504b4de0cddc6a2ef838b3037d12d318f18e665bf1a68602182e62", "retrieved_at": "2026-10-16T10:37:07+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "shyamsfo/bernoulli", "url": "https://github.com/shyamsfo/bernoulli", "category": "Projects", "topic": "implementation-repository", "description": "Self-hosted typed-decision implementation built on open-weight vision-language backbones, with HTTP, command-line and benchmark adapters; independent of Jev internals.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "83678bd5070bdab3693c58cc070ba23ba229f68efdd5719fa8c31f6dadf0917f", "retrieved_at": "2026-10-16T10:37:07+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "StefH/typesafe-ai.sdk", "url": "https://github.com/stefh/typesafe-ai.sdk", "category": "Projects", "topic": "implementation-repository", "description": "Community .NET SDK for the TypeSafe AI decision API.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b72f2c2de7b4b21db98df8f78bf80ebdedcbc92f428c222df3c4922a16417f2d", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "stern9/jev-bench", "url": "https://github.com/stern9/jev-bench", "category": "Projects", "topic": "implementation-repository", "description": "Benchmark harness comparing Jev with general language models on routing, binary checks and ordered ratings.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e3d2703c278f21c50856eb4644f5b603e3c413f1968c6a385aa7a370584c71ad", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "suharvest/laya-on-rk3588-jetson", "url": "https://github.com/suharvest/laya-on-rk3588-jetson", "category": "Projects", "topic": "implementation-repository", "description": "Laya and Kev deployment recipes for embedded hardware, with documented latency measurements and acceptance checks.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "7bac9db5744d293385aa49546ef75607362c76e4c5292c631a5c622a636b0a3c", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "SunixLiu/decision-datagen", "url": "https://github.com/sunixliu/decision-datagen", "category": "Projects", "topic": "implementation-repository", "description": "Policy-document data generator producing typed questions, rule-derived answers and teacher probability labels for decision-model training.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "451940d6367da4582d3fb913542f9fa8a63093d30a21ca16a5595789a83aded4", "retrieved_at": "2026-10-16T10:05:37+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Sweet-Butters/korea-ai-contest-tracker", "url": "https://github.com/sweet-butters/korea-ai-contest-tracker", "category": "Projects", "topic": "implementation-repository", "description": "Competition-directory application using Jev to classify discovered AI events and funding opportunities.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a0b4b99615059606af82a9612c1f1557bba7b86612b79230f5838b33f9eb5a4a", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "sxeptical/larp-detector", "url": "https://github.com/sxeptical/larp-detector", "category": "Projects", "topic": "implementation-repository", "description": "Browser extension using Jev to classify social-feed posts according to its stated persona-detection rubric.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ecf1dd52e4ed650fda479b641b1de0b3bfa8937babc1bcc62fe4faa849bfba92", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "sys1-bench", "url": "https://github.com/ugly-custard/sys1-bench", "category": "Projects", "topic": "option-order-evaluation", "description": "Decision-encoder robustness tests on a single-annotator core and a partly checked, LLM-assisted extension; examines option order, label naming and equivalent binary questions.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b8607b1d47266d815152684b240bc71ccf1fd0699f2611e1f98f767f14144220", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Sys1Cal", "url": "https://github.com/little-g-ai/sys1cal-v1", "category": "Projects", "topic": "evaluation-harness", "description": "Generators and saved reports for probability fidelity across typed primitives; exact targets are procedural, not human annotation of natural-language cases.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ddaf0646689ca54533acd7b20b5c9f5399d93c2d288eef2a55c430e90a27ab54", "retrieved_at": "2026-10-16T10:39:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "System One Security Evaluation", "url": "https://github.com/yxsec/system-one-security-eval", "category": "Projects", "topic": "evaluation-harness", "description": "Security-decision runners and saved-observation analyses; raw benchmark inputs and raw response text are not redistributed.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "8006bbcf86f24a673ef8dce91cf0e34ef52fe4aa6374f594084b7a0efb722b5b", "retrieved_at": "2026-10-16T10:39:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "System1Bench: Benchmarking Jev-Style Decision Models", "url": "https://github.com/cymcharming/system1bench", "category": "Projects", "topic": "decision-robustness-benchmark", "description": "Multi-suite decision evaluation with bilingual tasks, out-of-scope detection and option-order perturbations, keeping human-source and synthetic-reference results separate.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "fbdf0f17f94a6ad5fc8ae133448a694528c5744592f4ef3072aadc11840c3a8e", "retrieved_at": "2026-10-16T10:37:59+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "systemonedev/systemone-builder", "url": "https://github.com/systemonedev/systemone-builder", "category": "Projects", "topic": "implementation-repository", "description": "Toolkit for training, packaging, serving and evaluating typed-decision models, including Kenning, through a System One-compatible interface.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ae61b6b39be083337f6610effa47a6d254768ee4c11274a3642057cc7cccf2a1", "retrieved_at": "2026-10-16T10:37:08+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "tak-bro/local-jev-bench", "url": "https://github.com/tak-bro/local-jev-bench", "category": "Projects", "topic": "implementation-repository", "description": "Apple Silicon benchmark harness for local System One models, including multilingual accuracy and option-order sensitivity tests.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "581995ec84fbfe7f232aef2684a00c8f49c995c5c4176af277b24916fbaf3121", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Teddy9710/jev-playground", "url": "https://github.com/teddy9710/jev-playground", "category": "Projects", "topic": "implementation-repository", "description": "Local Jev workbench for composing typed questions and importing public benchmark cases.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "aa23d4212578de1facff64d54aa30749e48763e78461c5ab7793cb1011938339", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Temikus/lgtm", "url": "https://github.com/temikus/lgtm", "category": "Projects", "topic": "implementation-repository", "description": "Prerelease pull-request scoring tool combines typed model judgments into verdicts with a human-review path and configurable thresholds.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "73a82b9fb96ddab8b4e3464454186ad44cb20382756a0e1f19e4f0e38cf61889", "retrieved_at": "2026-10-16T10:37:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "TexasOct/jev-gateway", "url": "https://github.com/texasoct/jev-gateway", "category": "Projects", "topic": "implementation-repository", "description": "Session-aware model gateway routing chat requests using configurable decision rules and Jev integration.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "522ade8a60451acd8908badca2f1ca4b11709c2baa600d04ad2fdb119366be77", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "thiago-lage/jev-claud", "url": "https://github.com/thiago-lage/jev-claud", "category": "Projects", "topic": "implementation-repository", "description": "Application pattern separating typed Jev decisions from language-model prose, with offline mocked-provider tests.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ef9ff5a36f8e5d0e26e2bade955c5765bcf5eb52fba11b143bcbcfb3e468b514", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "thinkany-ai/autojev", "url": "https://github.com/thinkany-ai/autojev", "category": "Projects", "topic": "implementation-repository", "description": "Agent model router using Jev candidate selection and a local fallback when the decision provider is unavailable.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "1d653083e9d786462ab16b1046d84050125d2fee8c8f470bdc6a0eac50844368", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ThinkFlowLab/system1-agents", "url": "https://github.com/thinkflowlab/system1-agents", "category": "Projects", "topic": "implementation-repository", "description": "Agent examples applying System One models to browser actions, computer use, games and robotics.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "a5836edea1115fdb9d280e0958f528f99db9f454e0270c408a15579922d3bbb7", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "thusinh1969/BrighTO_Router", "url": "https://github.com/thusinh1969/brighto_router", "category": "Projects", "topic": "implementation-repository", "description": "Rust gateway combining model load balancing, fallback routing and System One-compatible provider endpoints.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "6e5d23d6117ee3b0f6ecd2958ea2a5d5cb849b742257d8014d99298a5dd65300", "retrieved_at": "2026-10-16T10:05:36+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "tnromy/evaluating-system-one-models-for-indonesian-regional-languages", "url": "https://github.com/tnromy/evaluating-system-one-models-for-indonesian-regional-languages", "category": "Projects", "topic": "implementation-repository", "description": "Preregistered evaluation of typed models and classifier baselines on sentiment and emotion in Indonesian regional languages.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "2a85b7d0ea8a92cf2a915e0a080efa89780c8e3165068b7c7f3d8e5898cb2d67", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "trungnt13/claude-system1", "url": "https://github.com/trungnt13/claude-system1", "category": "Projects", "topic": "implementation-repository", "description": "Claude Code plugin using Jev or OpenAI decision APIs to choose reasoning effort.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d95a9f465c4ab79813e6cb79750adb29573e0ef47bb57bf6449ef6b01c2609e2", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "unbraind/pm-jev", "url": "https://github.com/unbraind/pm-jev", "category": "Projects", "topic": "implementation-repository", "description": "Local-first task-item decision tool with an optional hosted TypeSafe provider.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e92ddb00003c19df8c9cc3cfc5343ccc898d3c504327f4e9bbbdfe5fd1bbec14", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "vilinski/jev", "url": "https://github.com/vilinski/jev", "category": "Projects", "topic": "implementation-repository", "description": "Command-line adapter sends a common typed-question format to TypeSafe, Cloudflare, local Ollama and token-log-probability backends.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "3136427e6ce61268502f4d313d968ca3c778efc00e82af7448e783dd16d0f74e", "retrieved_at": "2026-10-16T10:37:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Viperwow/jeff", "url": "https://github.com/viperwow/jeff", "category": "Projects", "topic": "implementation-repository", "description": "Unified typed-decision API fronting Jev-compatible model providers.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "0d101fd7b34d900d631a1bddae7b5946859acb7eb0a02171d97cc295e7093a82", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Visual Jev — shared-image decision implementation", "url": "https://github.com/guanxuyu-sv/visual-jev", "category": "Projects", "topic": "implementation-repository", "description": "Shared-image decision code and scored outputs; source images and raw per-example predictions are omitted, while fresh predictions require upstream datasets, model weights and compute.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "25d66790bc28cf10f3bf2af532771cfbf9e16d4aecd8f693a9cc46ff4c96983a", "retrieved_at": "2026-10-16T12:21:32.279565+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "wingedsheep/argentum-engine", "url": "https://github.com/wingedsheep/argentum-engine", "category": "Projects", "topic": "implementation-repository", "description": "Game-rules platform with a development-mode Jev opponent choosing legal actions from engine-provided options.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "87862d6057366246bc0b6b03234f8d47909006e9c84843b40abe231447f413bd", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "xingqiux/jev-playground", "url": "https://github.com/xingqiux/jev-playground", "category": "Projects", "topic": "implementation-repository", "description": "Unofficial local Jev testing interface for typed questions, including documented experimental vision support.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "d3b6c5ee4cfc58b2af5939a0757ab1095cd7d64b4980f13d7cff3ef28ee18ae9", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "Xm-Coding/jev-pr-classifier", "url": "https://github.com/xm-coding/jev-pr-classifier", "category": "Projects", "topic": "implementation-repository", "description": "Pull-request classifier using Jev to assign code changes to configured categories.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "9c249439b43a1205e5ffc12a5f42060e66192fd7ec1083e3311c4dae4a33af21", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "xpressabhi/jev-browser", "url": "https://github.com/xpressabhi/jev-browser", "category": "Projects", "topic": "implementation-repository", "description": "Browser-agent component separating Jev action decisions from the browser harness that executes them.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "ffe1791f1681b6c1f0be0fb2d8746e1e7de24642a4e99aecbac46862ebe59df0", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "XYenon/ajevt-browser", "url": "https://github.com/xyenon/ajevt-browser", "category": "Projects", "topic": "implementation-repository", "description": "Bounded Jev browser tool integrating action selection with coding-agent and MCP interfaces.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "56f945e5ea16c8d68ae1119ca5162f90aded1632adb046619c06a0fff15b40a0", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "yuan-phd/jevmark", "url": "https://github.com/yuan-phd/jevmark", "category": "Projects", "topic": "implementation-repository", "description": "Independent Qwen/LoRA training project compares supervised learning with RLCD using decision-probability and calibration evaluations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "58535c3cd6e1e6f2c5445ecb2f33f54be6bc620830d6be753bcfda65dc2150e3", "retrieved_at": "2026-10-16T10:37:11+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "ZeroTang05/Situation-Puzzle", "url": "https://github.com/zerotang05/situation-puzzle", "category": "Projects", "topic": "implementation-repository", "description": "Situation-puzzle game using Jev for yes/no judgments, hints and assessment of proposed explanations.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "e7f4adcef5b86d4e8fa1e756274b139a0af0016e80632877db8ab48eb9a17fb7", "retrieved_at": "2026-10-16T10:05:39+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "zhairy/dsh-agent-swarm", "url": "https://github.com/zhairy/dsh-agent-swarm", "category": "Projects", "topic": "implementation-repository", "description": "Multi-agent harness plugin combining shared Jev judgment tools with task planning, evidence checks and provider fallback.", "relation": "direct", "verification_level": "content_screened", "content_sha256": "b297d3aedfab9d6ba09f74bfb0e6be847846dd1c1a32335cd6138246b7285a8b", "retrieved_at": "2026-10-16T10:05:38+00:00", "ledger": "ecosystem-sources.jsonl" }, { "title": "$C^3$: Confidence Calibration Model Cascade for Inference-Efficient Cross-Lingual Natural Language Understanding", "url": "https://arxiv.org/abs/2402.15991", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d50e109f6ce72dcd7a0c43d2ee9afb26de294fa545288dd4ac681074bffd84b6", "retrieved_at": "2026-10-16T10:03:05+00:00", "ledger": "research-sources.jsonl" }, { "title": "98$\\times$ Faster LLM Routing Without a Dedicated GPU: Flash Attention, Prompt Compression, and Near-Streaming for the vLLM Semantic Router", "url": "https://arxiv.org/abs/2603.12646", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a4ebb24879fb35fac431dac3c4a03903f604e8c5dce2679559f556c0f2aae8a6", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "A CaLLM Judge is a Better Judge: Mitigating Bias in LLM-as-a-judge", "url": "https://arxiv.org/abs/2610.18108", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "76a77de8b2c8578f8f0452eb2139cde86203fbc7628ffedbe56845429ca57fe5", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "A Close Look into the Calibration of Pre-trained Language Models", "url": "https://arxiv.org/abs/2211.00151", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e28b539d974d2320e50532739dc3b9e109c706179af627560e61bd0d4b5c248e", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "A Geometric Abstention Mechanism for Language Models", "url": "https://doi.org/10.1109/icsc67292.2026.00056", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b9569e219265f70cdd538990ccbb13ba409037b2374075e928d92de6933b96c2", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "A Multi-Criteria Decision Framework for Enterprise LLM Routing", "url": "https://doi.org/10.3390/info17060539", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "7b16778b13e20487390491eb0c162b6d856e6a91feefeda0c6e7e89374771342", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "A risk bound for ensemble classification with a reject option", "url": "https://doi.org/10.1109/ssp.2011.5967817", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "9aae07d6b04b1391afa2bd139620d67f4bbce0a15601005e31b09a62cc169940", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "A Statistical Framework for Auditing Behavioral Dependence and Induced Bias in LLM Judges", "url": "https://arxiv.org/abs/2604.07650", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "adc8a33414846bd890e4ddd4960f2254395ae03a48dab88baf45ba7cb560bd24", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Accelerating Pre-trained Language Models via Calibrated Cascade.", "url": "https://arxiv.org/abs/2012.14682", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "64617d7e8ea1559a4f319f7c2ea47b91ee43fcf9a0316b266f6591eec57da070", "retrieved_at": "2026-10-16T10:03:05+00:00", "ledger": "research-sources.jsonl" }, { "title": "Accuracy-Rejection Curves (ARCs) for Comparing Classification Methods with a Reject Option", "url": "https://horizon.documentation.ird.fr/exl-doc/pleins_textes/divers17-10/010049968.pdf", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ab854979a3aab74ea7c1a142130189bd5bd0044cd1737c8631b62b5276c0fa36", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Active Learning for Binary Classification with Abstention", "url": "https://arxiv.org/abs/1906.00303", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "fa70c1e9fc071a693055654cc437c8d9c7db528889f9224437a5eb46e4434c26", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "Adaptive LLM Routing for Scientific Workflows: Predicting Query Complexity to Optimize Cost", "url": "https://doi.org/10.1109/ipccc66453.2025.11304643", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cdd412216e1ffe0d2445449df1166db307be0d77a828523afacb8cb4478f2898", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Adaptive Platt Scaling with Causal Interpretations for Self-Reflective Language Model Uncertainty Estimates", "url": "https://doi.org/10.18653/v1/2025.findings-emnlp.999", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d02a645a3f2542776309a0449ab067bc7ebda342280375bc71a141323ebca09a", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "Adaptive temperature scaling for Robust calibration of deep neural networks", "url": "https://arxiv.org/abs/2208.00461", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "395e45eb93e2dd01871e542ba0b85709b68ff93b3ee69537af0caedfa2b0a76a", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "Adversarial Robustness in Text Classification through Semantic Calibration with Large Language Models", "url": "https://doi.org/10.1109/cnml68938.2026.11452433", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "78297d1b911523e988919147377b02f8e525181ffe5e4244c55bd92f0b41aa2f", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "ADVICE: Answer-Dependent Verbalized Confidence Estimation", "url": "https://arxiv.org/abs/2510.10913", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "01cb1f475bb72f9e1405fd02faec455e1691728051ec066d5daaedafdde9f3a5", "retrieved_at": "2026-10-16T10:00:59+00:00", "ledger": "research-sources.jsonl" }, { "title": "Agentic Routing: The Harness-Native Data Flywheel", "url": "https://arxiv.org/abs/2607.11399", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "97d040b28157125f9e5be9cb924bf32b5fdda0119eca902418570c6665a54585", "retrieved_at": "2026-10-16T10:01:25+00:00", "ledger": "research-sources.jsonl" }, { "title": "Agreement by Construction: A Deterministic Filter Inflates Inter-Judge Reliability in LLM-Judge Evaluation", "url": "https://doi.org/10.58425/ajt.v5i12.624", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "10f384f2ffe1a57a8bd12566c92183b855c02b430066ced7fc284cc5c36c3aa9", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "Alleviating Hallucinations from Knowledge Misalignment in Large Language Models via Selective Abstention Learning", "url": "https://doi.org/10.18653/v1/2025.acl-long.1199", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "273073abc29450f13a19864ce5db5fcbd7608106534a8bc511c870d8888d5ee8", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "Ambiguity Detection and Uncertainty Calibration for Question Answering with Large Language Models", "url": "https://doi.org/10.18653/v1/2025.trustnlp-main.4", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4ceb176275f4bed597a5ed87abdecc28e7190b16b44ad1cf62199b1947ec5191", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "An analysis of binary isotonic regression: degrees of freedom and implications for calibration", "url": "https://arxiv.org/abs/2607.27301", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f6167b4cef19e310d2e9824c3bbc88f9dd96113079340c74de5efaf598c36698", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "An Empirical Study of Security Calibration in Large Language Models for Code", "url": "https://arxiv.org/abs/2606.31159", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4ee83eef0ccbc720476766e36bcc44756642782c9ec78158cfa0369a9af8b81e", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "An Exploration of Multicalibration Uniform Convergence Bounds", "url": "https://arxiv.org/abs/2202.04530", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b81c1979539bb662241aecfe4501c4ae4ffdcd2c56aa691016fa03f511471f93", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "An Ordinal Data Method for the Classification with Reject Option", "url": "https://doi.org/10.1109/icmla.2009.11", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0477eb1f407b09f85ed3ccd5ad77f341b745700be0db067b23b8144a6150b743", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "ANALYSIS OF CLASSIFICATION WITH A REJECT OPTION", "url": "https://doi.org/10.1142/s0219691309002970", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "525354ab6fc52c015311596c06fefec3d1db0db8a84d1a01f495311d8b443a0b", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Analysis of LLM-as-a-Judge Position Bias Mechanism Based on Controlled Synthesis Data", "url": "https://doi.org/10.54254/2753-8818/2026.33416", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "760087524d7f8c6cbd786cebd7b52674dc1850346c40e65484a4f4572f56c630", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Anchoring Bias in LLM-as-a-Judge Systems: Prior Scores Compromise Evaluation Independence", "url": "https://arxiv.org/abs/2608.25869", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c580c359f210ac2409e29b0b447b6b09eefd959deb46b78fb9c2913403f5786a", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Another look at combining rejection-based pattern classifiers", "url": "https://doi.org/10.1109/icpr.2000.906038", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "8abf86ac040a4adb5907e78c6d747aff2d6fd5cc3722ce8a997e61deddeb6fe5", "retrieved_at": "2026-10-16T10:02:45+00:00", "ledger": "research-sources.jsonl" }, { "title": "Assistant-Guided Mitigation of Teacher Preference Bias in LLM-as-a-Judge", "url": "https://arxiv.org/abs/2505.19176", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0f7d69bf34eccc0670b1bfe674347b55533de21869d0c4c353f63315a0b01b63", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Asymptotic Risk Calibration for Selective Question Answering", "url": "https://arxiv.org/abs/2608.12008", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "248ef5c3ac84f3f1ab9d4cebc51b500187daf9b3fb324bf2de2c647ba30e332e", "retrieved_at": "2026-10-16T10:02:55+00:00", "ledger": "research-sources.jsonl" }, { "title": "Attention Meets Reachability: Structural Equivalence and Efficiency in Grammar-Constrained LLM Decoding", "url": "https://arxiv.org/abs/2603.05540", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1046e00326ea96060399332a32880228f389d9be83be150fddc5bd6297428ccc", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "Auto-Rubric: Learning to Extract Generalizable Criteria for Reward Modeling", "url": "https://arxiv.org/abs/2510.17314", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "eabc23cbdced87a4f2fdeffa9527efafc7d31787ba822416f913e38ba94b63fd", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Balancing Classification and Calibration Performance in Decision-Making LLMs via Calibration Aware Reinforcement Learning", "url": "https://doi.org/10.18653/v1/2026.findings-acl.610", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f946772a2009f48c1b3abe5778befe5531c6d62548af49bc84c65204fc07e678", "retrieved_at": "2026-10-16T10:02:42+00:00", "ledger": "research-sources.jsonl" }, { "title": "Balancing Two Classifiers via A Simplex ETF Structure for Model Calibration", "url": "https://arxiv.org/abs/2504.10007", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c454b83d9a80e1fe1f653661bfa62d4da7bc7ac31bd314d3be61aa0c049a900c", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Benchmarking Goodness-of-Fit and Calibration Algorithms for Logistic Regression Classifiers: A Large-Scale Simulation Study under Sparse Data", "url": "https://arxiv.org/abs/2607.16344", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c87295186323919509f9ac2f8c29086d342c9fe768f67875aab3a1ed11591ed4", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Benchmarking LLM Judges for Voice-Agent Evaluation: Reliability, Calibration, and Human Oversight", "url": "https://arxiv.org/abs/2608.24314", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "51e8646f8af0b9fc6bfc1441b4deeabf472a40c25e0fd5eab3f18afca63d181c", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beta calibration:a well-founded and easily implemented improvement on logistic calibration for binary classifiers", "url": "https://hdl.handle.net/1983/6e0c0e3e-9911-4be8-85dd-27358e23c98c", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "63e0ca990c5a96feb6d97abe674a9ac2d53f89f9e35de5873d7c1bab55ccece0", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Better Classifier Calibration for Small Datasets", "url": "https://arxiv.org/abs/2002.10199", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1b89ccee62c6bb83d0711c704027aee0799ac8e2dedd1fd3da75fdc36bfaa4e3", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beyond Accuracy: The Role of Calibration in Self-Improving Large Language Models", "url": "https://arxiv.org/abs/2504.02902", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e830bcf2d636423ab2c3db202605299330b291e1e4eb6d07f6ab6ea160ee3228", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beyond Consensus: Downward Bias and Role Asymmetry in Multi-Agent LLM Judges for Subjective Evaluation", "url": "https://arxiv.org/abs/2608.30373", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cd09b55388fe8ebd8948b10169029878bf85431d378968bca1231f67148ac25e", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beyond Consensus: Mitigating the Agreeableness Bias in LLM Judge Evaluations", "url": "https://arxiv.org/abs/2510.11822", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "87554a0a89f113d514108247513d0761872d2163265683561216ce8cbf4acd18", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beyond Gemini-3-Pro: Revisiting LLM Routing and Aggregation at Scale", "url": "https://arxiv.org/abs/2601.01330", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "be56abf4918d399a6037334f5281f722db6362be72fd0165a1a5100dcfb3f028", "retrieved_at": "2026-10-16T10:01:22+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beyond Static Benchmarks: A Validity, Reliability, and Sociotechnical Framework for Evaluating LLMs in Deployment Contexts", "url": "https://doi.org/10.18653/v1/2026.evaleval-1.30", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e53408b897a354d64db3053aaea499a0263ffab3729bc15535d5a7a398bfde31", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beyond the Final Layer: Intermediate Representations for Better Multilingual Calibration in Large Language Models", "url": "https://arxiv.org/abs/2510.03136", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "58250bea1dca8f06d6a134dca1452479755b608a81eeb76ebfbe83de678b685f", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Beyond the Surface: Enhancing LLM-as-a-Judge Alignment with Human via Internal Representations", "url": "https://arxiv.org/abs/2508.03550", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "04ecdfd41ba81fbe7f3f04f1b6bf12ac2e8270424ace723fd1eab860c3ef2cea", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Bias and Uncertainty in LLM-as-a-Judge Estimation", "url": "https://arxiv.org/abs/2605.06939", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d08d78d556e2e6a5c5abdf849722b504cba0e01999d37e22d62d010a0a5869f8", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Binary Classification with Bounded Abstention Rate", "url": "https://arxiv.org/abs/1905.09561", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e6e868aaad9de19a1df107d4fd1720cc4661fa82981af234e5120891d7bf22a2", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "Binary Classifier Calibration Using an Ensemble of Near Isotonic Regression Models", "url": "https://arxiv.org/abs/1511.05191", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "84b4c7e6af67a29260510ac35fe7a4d21cf7bfaa8d097df1b82395ffeac713c9", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "Binary Classifier Calibration: Bayesian Non-Parametric Approach", "url": "https://arxiv.org/abs/1401.2955", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3ca2044ef6594b823ee48680258ca0ba7bcfdedb5faa941fa534d3b5c9a829c8", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Binary Classifier Calibration: Non-parametric approach", "url": "https://arxiv.org/abs/1401.3390", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "db0ce09e9339c7fb01af0cfeef64dc4e87759e33b1d62f90b619a6e5a611aa68", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Breaking Model Lock-in: Cost-Efficient Zero-Shot LLM Routing via a Universal Latent Space", "url": "https://doi.org/10.1609/aaai.v40i43.40970", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "663b24110173b2276f9c532469f0e905a979ca214cd52501b814e716de5f08b1", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Budget-Aware Agentic Routing via Boundary-Guided Training", "url": "https://arxiv.org/abs/2602.21227", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "778765adca66ad1625d53fd231c7789dae37239f632bc6095531f2d2c35d894f", "retrieved_at": "2026-10-16T10:01:21+00:00", "ledger": "research-sources.jsonl" }, { "title": "C3PO: Optimized Large Language Model Cascades with Probabilistic Cost Constraints for Reasoning", "url": "https://arxiv.org/abs/2511.07396", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6ffdede828af3cd02f39992f663efee48f3951a86acb17e7224e0793c09968e5", "retrieved_at": "2026-10-16T10:03:05+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calculating classifier calibration performance with a custom modification of Weka", "url": "https://doi.org/10.1063/1.4907827", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "85796c44ec2c5ea1b7f8f0fe0d764bab0387036f5f4d983ad205656686107ed8", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibrated Confidence Estimation for Tabular Question Answering", "url": "https://arxiv.org/abs/2604.12491", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "9bdc828fed5fcd30c49fe851f6496ed9caea7323542dec33180b041cb050955b", "retrieved_at": "2026-10-16T10:01:00+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibrated Learning to Defer with One-vs-All Classifiers", "url": "https://arxiv.org/abs/2202.03673", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "18221a31bead7c98c6a636347c0665c20ef0c98ed8b6e0d6e1cf3ce4b984356b", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibration of Neural Networks", "url": "https://arxiv.org/abs/2303.10761", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "78c6814b1e292448744467a52d3f135ffe26adec5dacfd084bfdd822857fdd3d", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibration of Neural Networks using Splines", "url": "https://arxiv.org/abs/2006.12800", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ae0cd0f24cca83162560ffd14db7d1e0b72c2dcd684ca3d604c468d277d9df16", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibration Scoring Rules for Practical Prediction Training", "url": "https://arxiv.org/abs/1808.07501", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2c4870b6baea7b9fd194a2bd65f5904e08ee857a2ab0be0ef30c9eea00f0e95e", "retrieved_at": "2026-10-16T10:02:42+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibration, Coherence, and Scoring Rules", "url": "https://doi.org/10.1086/289244", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f0fa9cf66f6c772c649549122b05a3408b0f260a7a22d458fef2cc63e480b0ee", "retrieved_at": "2026-10-16T10:02:42+00:00", "ledger": "research-sources.jsonl" }, { "title": "Calibration, Entropy Rates, and Memory in Language Models", "url": "https://arxiv.org/abs/1906.05664", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a6a3018af436b56955fddf025d52a019b4c6ba3ffe91298fe04e9d792f637492", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "CalVerT: Augmenting Agents with Calibrated Verifier Telemetry Improves Action and Learning in Knowledge-Intensive Tasks", "url": "https://arxiv.org/abs/2606.21777", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "812c10f2db9355f6b0664d50d0e6ea932d078eb73811062501774a6fc8c778bc", "retrieved_at": "2026-10-16T10:01:01+00:00", "ledger": "research-sources.jsonl" }, { "title": "Can LLM-as-a-Judge Reliably Verify Rubrics in Agentic Scenarios?", "url": "https://arxiv.org/abs/2606.29920", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3ca87fc1482d0ca2644ad19aaf7dc913dfd41fa9091f461ce5a33957e6a83b5c", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Can LLMs Replace Human Evaluators? An Empirical Study of LLM-as-a-Judge in Software Engineering", "url": "https://arxiv.org/abs/2502.06193", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e1b6b2cff4e395a31cdf215f20f5b6d6de9d6bcc7314f20ea191fc8d5c97e4ab", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Can We Trust LLM Judges: A Study of Capability-Dependent Biases and Multi-Judge Ensemble for Bias Calibration", "url": "https://arxiv.org/abs/2609.12002", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "20e959a3dcf8364fe33c7ef43b6f7a4e53d5176f45daa0aeea75507fe2a580e9", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Can We Trust This Evaluation? Instance-Level Reliability of LLM-as-a-Judge", "url": "https://arxiv.org/abs/2610.14029", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4e8d39c7c26c65b430d17fdc8f967d649369b29ecda8263a9cf2bc22c635d014", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "CARROT: A Cost Aware Rate Optimal Router", "url": "https://arxiv.org/abs/2502.03261", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5a72f9563f25942e0c2618baecdc97e1c7a832e5c98ee848c71a3d3037ce6181", "retrieved_at": "2026-10-16T10:01:21+00:00", "ledger": "research-sources.jsonl" }, { "title": "Cascaded Language Models for Cost-Effective Human–AI Decision-Making", "url": "https://arxiv.org/abs/2506.11887", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "656d8ef7fbecae793f08d0e6e3eaad4776859c9ac94c6dacecfbf4463ba2075f", "retrieved_at": "2026-10-16T10:03:05+00:00", "ledger": "research-sources.jsonl" }, { "title": "Cascaded Neurosymbolic Code Generation for Niche DSLs: Preserving Chain-of-Thought in Grammar-Constrained Decoding", "url": "https://doi.org/10.3390/software5030033", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a38e354f186c9f78bb279955d00da1114e3b8bec77edd88ce0e83d9b74a4326c", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "CDRRM: Contrast-Driven Rubric Generation for Reliable and Interpretable Reward Modeling", "url": "https://arxiv.org/abs/2603.08035", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5e3e18f2c8f030755667b4f9284a706f36a0b2b5a8be1ba7c2b6c39ef7cd339e", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Chain-of-Models: Cross-Model Auditing for Bias-Robust LLM Judges", "url": "https://arxiv.org/abs/2607.28636", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0f3e3d66225f046359ac15b7dcf9eafacd17db68f80b73db8912386c40ea7e8b", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "CHASE: Channel-Wise and Spatial Attention for Early Exiting in Image Classification", "url": "https://doi.org/10.1109/icassp49660.2025.10888714", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cdff9bcf34359fb15a7bb36d459af495457ac6ab81a97a4abd9ce424ff3294c9", "retrieved_at": "2026-10-16T10:03:10+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification Under Partial Reject Options", "url": "https://arxiv.org/abs/2202.14011", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f779bb28c46da451830e3a8456a44e4dd6acbcf8d445b6f154f5e02c1ad7258c", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification with a reject option under Concept Drift: The Droplets algorithm", "url": "https://doi.org/10.1109/dsaa.2015.7344808", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1e1c19e0018adc713c3725140bb9c3615e2a09c40bcdc1c91fb5b15b20954364", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification with a Reject Option using a Hinge Loss", "url": "https://doi.org/10.5555/1390681.1442792", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a956a605bbcd2d2df841903cdc1967959d4c9bf6ce9c1842c72f42f502176327", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification with abstention but without disparities", "url": "https://arxiv.org/abs/2102.12258", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "fc0986f6a25b444029e2dad07ecc80a8d4d98ba7392b315b82f28a42f2b05f14", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification with Abstention Under Class-Conditional Error Constraints", "url": "https://arxiv.org/abs/2609.22632", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "27fc4058c80c579e61d4f0505c3b305a49be3e8273fd595a4d24607454945991", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification with reject option", "url": "https://doi.org/10.1002/cjs.5550340410", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f76aa29ddfb2c70920ee483b6d63a41e2381e81a3e7e84ca12a721f49ead0efe", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification with reject option in text categorisation systems", "url": "https://doi.org/10.1109/iciap.2003.1234113", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1e7847016a52cc070080b8623210e1c7b27d6fc0141edf109669e8c9e373dcaf", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classification with reject option: Distribution-free error guarantees via conformal prediction", "url": "https://arxiv.org/abs/2506.21802", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cfe32bfc3869f80cf1e907cd2d59c92eb4fc7cebaae0d40ba6d0e72d3e925781", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classifier Calibration at Scale: An Empirical Study of Model-Agnostic Post-Hoc Methods", "url": "https://arxiv.org/abs/2601.19944", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e7778739e41738406c89190f5c6c0d066ed0a46c8d173f9133685f00e37a89cf", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classifier Calibration with ROC-Regularized Isotonic Regression", "url": "https://arxiv.org/abs/2311.12436", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a00deed3b1546a04fd6aa052370fd325deab79f4d39730acb93593321cea4857", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classifier Calibration: With Application to Threat Scores in Cybersecurity", "url": "https://arxiv.org/abs/2102.05143", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f70b8280f8c5e1959aae3ced40a81c2078c39ffd0cc5ad7d8bf372e47998f58f", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classifier Ensemble for Efficient Uncertainty Calibration of Deep Neural Networks for Image Classification", "url": "https://arxiv.org/abs/2501.10089", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "decfddb2107224daf2ac445a45efa951825f33f77f42757e96be2f4541d268a9", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Classifiers with a reject option for early time-series classification", "url": "https://doi.org/10.1109/ciel.2013.6613134", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "057a29c8ce5e89a98862142547be2b16dcaf5ddf9eced2bf91b8bf344b117e5e", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "CLEAR: Error Analysis via LLM-as-a-Judge Made Easy", "url": "https://arxiv.org/abs/2507.18392", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4a1fc612fa4dbfe443780560cc92b91fdbf992ded08fda2a6cd6cf4e068cc665", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Condition-Stratified Robustness Analysis of Post-Hoc Calibration Methods for Probabilistic Classifiers", "url": "https://arxiv.org/abs/2607.11542", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c9ec056889f80b6dac36c1e89438604f08549342c8eeda74bb1b8a4f41911eb2", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Confidence as a Tie-Breaker: Reassessing Multilingual Hedging Bias in LLM-as-a-Judge Evaluation", "url": "https://doi.org/10.18653/v1/2026.acl-srw.33", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c3725d15d76dbd8f6e3b6243e8ce3dbc5066581caaf607c40a3724f6c3363531", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Confidence Calibration in Large Language Models for Uncertainty Quantification: Affecting Calibration with Conditional Weight Updates", "url": "https://doi.org/10.1609/aaaiss.v7i1.36937", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ccc5eb98cff3685e257cd8ea3d7ec5a45db9c6ca4274b9fd6b24f3444089acfc", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Confidence Improves Self-Consistency in LLMs", "url": "https://arxiv.org/abs/2502.06233", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a9482a6d9bc3c79214cf8b627ea56772034cf83ccb3255e9270e4a81d564c8ee", "retrieved_at": "2026-10-16T10:00:59+00:00", "ledger": "research-sources.jsonl" }, { "title": "Conformal LLM Routing with Distribution-Free Safety Guarantees", "url": "https://doi.org/10.18653/v1/2026.acl-srw.70", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3374e45b7290b21c0911b40b2cb5c9ee53210b209e94327b3581a201d1d74668", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Consistent Algorithms for Multiclass Classification with a Reject Option", "url": "https://arxiv.org/abs/1505.04137", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cd61c3f08a2e0167cb888fc0e3dadd1eae89867709f9f532147e4db01a97f853", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Contamination-Free LLM Routing on LiveBench Reasoning Tasks: Accuracy-Cost-Latency Tradeoff Learning", "url": "https://doi.org/10.70382/ajsitr.v12i9.088", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c4397026d5a172324b0e088444d26d728ba4712d920e6b1dfce857e38a7d24d3", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Contrastive Cross-Lingual Calibration for Large Language Models", "url": "https://doi.org/10.64748/8gkysx36", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "16182cabe7a384cf80616b2d246eb60547cc780bb25a191f3e050d058ba207ed", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Controlled Abstention Neural Networks for Identifying Skillful Predictions for Classification Problems", "url": "https://arxiv.org/abs/2104.08281", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1e459334921035a9d12d1bb81e4b6842dda947d5bf2ecc8b463870916bc63585", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "Copy-as-Decode: Grammar-Constrained Parallel Prefill for LLM Editing", "url": "https://arxiv.org/abs/2604.18170", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "63d935d7e7a056295c57418120ec622040f843d0c15b35942b257da430fe585b", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "Correlation-Aware Contextual Bandits with Surrogate Rewards for LLM Routing", "url": "https://arxiv.org/abs/2607.09015", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "91932a1060cf357c446647fd590d30174227a16eeb7491f3c4c6b7682a8b6dda", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Cost-Aware Calibration of Classifiers", "url": "https://doi.org/10.1287/ijds.2024.0038", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ce148f0414651434583c08b37ee4f1e8cf1e7618610d0633ffeca09e498171cd", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Cost-Optimal LLM Routing with Limited User Feedback under User Satisfaction Guarantees", "url": "https://arxiv.org/abs/2606.19376", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "65389fb337baf9ab80969f5261858cf36bf5ef7f1bacec5b77720ce07318fae7", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Cost-Sensitive Learning to Defer to Multiple Experts with Workload Constraints", "url": "https://arxiv.org/abs/2403.06906", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d17265b6dcba9826253dd10c4fce2d200a427048504c525c53c46cbe43900385", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "CriterAlign: Criterion-Centric Rationale Alignment for Code Preference Judging", "url": "https://arxiv.org/abs/2605.19665", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "df62d358d274241728776f06932de8ab21a8c59fb72b012dc8e352ba30083806", "retrieved_at": "2026-10-16T10:01:17+00:00", "ledger": "research-sources.jsonl" }, { "title": "Cross-lingual Calibration of Pre-Generation Success Probes for Multilingual LLM Routing", "url": "https://arxiv.org/abs/2610.06216", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "34541d73de3c34c519b02fa89f1bc83bef26bafa952b1ed07487ed367ef553bc", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Crowd Comparative Reasoning: Unlocking Comprehensive Evaluations for LLM-as-a-Judge", "url": "https://arxiv.org/abs/2502.12501", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a0a8bb1b6c4d43dacc778b07937e6215d6fb02988a97e43be0387a8bd037caf0", "retrieved_at": "2026-10-16T10:01:17+00:00", "ledger": "research-sources.jsonl" }, { "title": "Curse of Knowledge: When Complex Evaluation Context Benefits yet Biases LLM Judges", "url": "https://arxiv.org/abs/2509.03419", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d021e05d4a1135f1ee8158b119d3ca6a5a59bf59d0287beeddcef99bc9178b3e", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "CyclicJudge: Mitigating Judge Bias Efficiently in LLM-based Evaluation", "url": "https://arxiv.org/abs/2603.01865", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "84094119a236cd19c6fcebc44eae87c5fc3e1ae9312f3a9783b227330f03bb8c", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "DBCal: Density Based Calibration of classifier predictions for uncertainty quantification", "url": "https://arxiv.org/abs/2204.00150", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "16e5489c93b6928cc97075b021c0b60519e12633532dd67e2cdec36195a8bbeb", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "DecoCal: Decoding with Calibration in Diffusion Large Language Models", "url": "https://doi.org/10.18653/v1/2026.acl-long.545", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2d61795bcb0d90a5f188420693cc6ba4b51898d29d12f8330a1d063481ec4314", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Decode-Time Grammars: Constrained LLM Generation over a Refinement Order of Grammar Fragments", "url": "https://arxiv.org/abs/2607.18357", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5fd50928d53ed36ede944d7ac7d272bdd5abaac3186531b542f86a2cdb094bf8", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "Decomposition Enhances Reasoning via Self-Evaluation Guided Decoding", "url": "https://arxiv.org/abs/2305.00633", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "aa70416de34167ca7b60c8d8417ef8ee306513573a48889e38dbad8d6a02e8a5", "retrieved_at": "2026-10-16T10:01:04+00:00", "ledger": "research-sources.jsonl" }, { "title": "Deferring Concept Bottleneck Models: Learning to Defer Interventions to Inaccurate Experts", "url": "https://arxiv.org/abs/2503.16199", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "129937ad841d7ce8b79015f3bf6efb5333309bb68ee034f36c4067a56d2e6189", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "DeltaRubric: Generative Multimodal Reward Modeling via Joint Planning and Verification", "url": "https://arxiv.org/abs/2605.09269", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6138c36945b3f3900e37aae407f427b81da9a076b6a29075acc0ab91ec965cb0", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Dirichlet-Based Prediction Calibration for Learning with Noisy Labels", "url": "https://arxiv.org/abs/2401.07062", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1f8ec4fe601b01630a9e8240e9b60b6efea7f14a123be253bf9aa948de5517db", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "Dirichlet-based Uncertainty Calibration for Active Domain Adaptation", "url": "https://arxiv.org/abs/2302.13824", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3312669fabe8c0399782563086974aca75951ba9dd82ea62554f35e9cd98d2a1", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "Distance-aware Calibration for Pre-trained Language Models", "url": "https://doi.org/10.18653/v1/2024.findings-emnlp.725", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "526b6cac77b1be2643d86a21b43bbe673f609f5dc98c8fda45db53e9b3fddffa", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Distilled Hierarchical Neural Ensembles with Adaptive Inference Cost", "url": "https://inria.hal.science/hal-02500660", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6245cc6263a3703972f28e79748f2a738000075c444793c46943be2456aee22a", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "Do MLLM Judges Judge the Edit? Auditing Bias in Image Editing Evaluation with Verified Quality Preservation", "url": "https://arxiv.org/abs/2610.01670", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "68a1fb6145f2891abfd776d44b554359a4542b181f6802a58ee39da8c047ea98", "retrieved_at": "2026-10-16T10:01:15+00:00", "ledger": "research-sources.jsonl" }, { "title": "Domain Recentering and Confidence-Weighted Prior Calibration for Vision-Language Models", "url": "https://arxiv.org/abs/2609.29358", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "78993b7592affddb1915e07b626c84ffcd4b317ab9c74dc253ad76634723db13", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Domain-Aware LLM Routing During Generation", "url": "https://doi.org/10.1109/bigdata62323.2024.10825152", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ab2a9f7d6301ef59be8641d14c7e6ff47173bcd4ed443bffffb7bbb9aac2d373", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Drift-Aware LLM Routing with Sparse Contexts and Shared Budgets", "url": "https://arxiv.org/abs/2609.00662", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "845dc5aff357988d4d50be1ab370f7185e0df85bee2c34decb2fc85412c4d15e", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Dual-Scale Fuzzy Boundary Abstention for Selective Prototype Classification", "url": "https://doi.org/10.54216/jnfs.110201", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d9226578f9c9e411ada59a28f2c70852523a948096443556fcbee6a7f1ff2630", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "Efficient Grammar-Constrained Decoding via Parser Stack Classification", "url": "https://arxiv.org/abs/2608.03065", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "77f22b046e792eeeafad51960aedd5cbcbf6e8c3e17f023851614f4e733ba8e0", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "Efficient Out-of-Scope Detection in Dialogue Systems via Uncertainty-Driven LLM Routing", "url": "https://arxiv.org/abs/2507.01541", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5261bca1f153a6905a1a9d93de772a9de1fb18f84a5e49e65cde03eb87ab3a3e", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "ELF: An Early-Exiting Framework for Long-Tailed Classification", "url": "https://arxiv.org/abs/2006.11979", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "036588c1356961fee0c2a617ac3b8c6e0fc0a9212952e05f3eb69b041753b968", "retrieved_at": "2026-10-16T10:03:10+00:00", "ledger": "research-sources.jsonl" }, { "title": "Elicitation of Probabilities Using Competitive Scoring Rules", "url": "https://doi.org/10.1287/deca.1030.0003", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c81d8ad19a1bd097e1f904a85c5bbfe12976c716b19e76fd35cd5b3166204409", "retrieved_at": "2026-10-16T10:02:40+00:00", "ledger": "research-sources.jsonl" }, { "title": "Eliciting forecasts from self-interested experts: scoring rules for decision makers", "url": "https://arxiv.org/abs/1106.2489", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ac3637233f2f9d517ef5655d3ee22412449341ec4fb8e456f8f5411ae40d2980", "retrieved_at": "2026-10-16T10:02:40+00:00", "ledger": "research-sources.jsonl" }, { "title": "Enabling Calibration In The Zero-Shot Inference of Large Vision-Language Models", "url": "https://arxiv.org/abs/2303.12748", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "28bb2be9714606df726a6a722fa128fe1e4ebdbc7cdef9f41b69f52e76fbfd59", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Enhancing Automatic Text Evaluation through Calibration-Aware Large Language Model Autoraters", "url": "https://doi.org/10.14293/pr2199.003506.v1", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "42a24cd6af4386b9cc5dbf81c22796f87fa2ac3fe1dd98979bb93fa2194ed045", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Enhancing Language Model Factuality via Activation-Based Confidence Calibration and Guided Decoding", "url": "https://arxiv.org/abs/2406.13230", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "8d0e5e7c24880171997870f51b1c017b8f3ae50cf44b1260da0d8cff84e7d822", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Entropy After for reasoning model early exiting", "url": "https://arxiv.org/abs/2509.26522", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "204f349d8ef65141b40d0a43e310846449fb11641a376e67cc3ef900ec3e5f2f", "retrieved_at": "2026-10-16T10:01:25+00:00", "ledger": "research-sources.jsonl" }, { "title": "Essays on probability elicitation scoring rules", "url": "https://doi.org/10.1063/1.4759598", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c343db630940e9b491be01a11d7168dce738da9763f891ccb1ea5809773a9bab", "retrieved_at": "2026-10-16T10:02:40+00:00", "ledger": "research-sources.jsonl" }, { "title": "Evaluating Escalation Signals for LLM Routing: Targets, Controls, and Five Ways to Fool Yourself", "url": "https://arxiv.org/abs/2610.07354", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "002fc753e3b0bfbd11ee0b4788985d09219f5c3858a99de9ec4c346ec093ec4b", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Evaluating Rubric Generation with Interventional Transfer", "url": "https://arxiv.org/abs/2610.10809", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1f82fed8a1cabcfdbf337ce84122d42ca99d3f16f28207c4cc54febb3bf6f654", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Evaluating the impact of prevalence on the scaled Brier score, Matthews correlation coefficient, and other performance metrics for binary and multiclass hard classifications", "url": "https://doi.org/10.1007/s44248-026-00119-w", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ad4cf395626b4ee8d41ddb752fc6fe7d83cc5e884518fa17321136c8a40732ad", "retrieved_at": "2026-10-16T10:02:41+00:00", "ledger": "research-sources.jsonl" }, { "title": "Evidential Rule Learning for Interpretable Classification with Abstention", "url": "https://arxiv.org/abs/2608.05859", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "71e2a13f03f7f5ea322cf755db76cd51363cdc11d5e82912e5bae8b977a359e8", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "Exp-Concavity of Proper Composite Losses", "url": "https://arxiv.org/abs/1805.07737", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "13319488eafb4bc639fb13afed00675d2f6f66600e21c8bf05c885e165a057aa", "retrieved_at": "2026-10-16T10:02:39+00:00", "ledger": "research-sources.jsonl" }, { "title": "Expectation consistency for calibration of neural networks", "url": "https://arxiv.org/abs/2303.02644", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "34fd5af005b85018cefc3c44ba35515c0d854c1414368589617964e8c00b0561", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "F-PABEE: Flexible-Patience-Based Early Exiting For Single-Label and Multi-Label Text Classification Tasks", "url": "https://arxiv.org/abs/2305.11916", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "8c3adf3d38004bb9f995be76a0d74fdf02d50cd3800fa1c6d41c09f11817985d", "retrieved_at": "2026-10-16T10:03:10+00:00", "ledger": "research-sources.jsonl" }, { "title": "Fairness or Fluency? An Investigation into Language Bias of Pairwise LLM-as-a-Judge", "url": "https://arxiv.org/abs/2601.13649", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e4df3ec7e429f813a231b7796e2712428af272cd0af186a7642f5d180e0b9208", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Fine-Tuning Large Language Models to Appropriately Abstain with Semantic Entropy", "url": "https://arxiv.org/abs/2410.17234", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "48fd3d9ddfe342dc48f13e056a084ea7ca893bf8fd687236bbf6def18502050c", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Fixed-Sequence Confidence-Interval Calibration for Finite-Sample Selective Risk Control", "url": "https://doi.org/10.3390/stats9050109", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "40abee34fbe6234e0a23e0a1b9952fa4ff9e70737351418157d5be17fd6f7625", "retrieved_at": "2026-10-16T10:02:55+00:00", "ledger": "research-sources.jsonl" }, { "title": "Focal Calibration Loss: Controlling Posterior Distortion in Deep Neural Classifiers", "url": "https://arxiv.org/abs/2410.18321", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f8492dc498f9207b425076edfc556f847b8fa6ae5d83ada32942beff4ee6a904", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "From Code to Courtroom: LLMs as the New Software Judges", "url": "https://arxiv.org/abs/2503.02246", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4cc12d25592546bb99b9fcd80cef398809e5359529de2fa4d5ce1e99cc302b98", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "From Deferral to Learning: Online In-Context Knowledge Distillation for LLM Cascades", "url": "https://arxiv.org/abs/2509.22984", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "dc6503fa4ac1cc36cf21a33e6a395a492ee1ad61e6231475a129b81a0d17f5a6", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "From Holistic Evaluation to Structured Criteria: Rubrics Across the Evolving LLM Landscape", "url": "https://arxiv.org/abs/2606.08625", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "8b6a839b239e11ab1ab9805b5037f15610b291dcda603885e8ca62ac9e5f2942", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "From Myopic Selection to Long-Horizon Awareness: Sequential LLM Routing for Multi-Turn Dialogue", "url": "https://arxiv.org/abs/2604.12385", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "38a48d9c874dafb2eda97a25aac04cef4deb2fe08250e1aa8f0bff74091b551c", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "From Rubrics to Recipe: Principle-Centric Benchmark for Evaluating Large Language Models", "url": "https://doi.org/10.18653/v1/2026.evaleval-1.15", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "632b687cff547f71ce74a6df5aad1864ec7d2a50423c4303a8f75e3e444978b5", "retrieved_at": "2026-10-16T10:02:57+00:00", "ledger": "research-sources.jsonl" }, { "title": "From Sampling to Cognition: Modeling Internal Cognitive Confidence in Language Models for Robust Uncertainty Calibration", "url": "https://doi.org/10.1609/aaai.v40i37.40424", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b02072cdfa5e4f8de07f6df042769ee4b103c7d248b7da386ed66bfd3520e930", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Full-ECE: A Metric For Token-level Calibration on Large Language Models", "url": "https://arxiv.org/abs/2406.11345", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "90fd7d48ce7b6172a4024fadaa6bc5322a7274227e362e0c6892564da99f80e4", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Full-range Binary Classifier Calibration for Stable Model Updates in Production", "url": "https://arxiv.org/abs/2607.05481", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "8ba562c50bf2f75b4f30bea8b91aa1c89c70de0e1b40f9b8b7e919bf39134021", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Fuzzy classification with reject options by fuzzy if-then rules", "url": "https://doi.org/10.1109/fuzzy.1998.686333", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "89418fa9d71732602ffa78ce3cd29e956483b366430612c528304bc48fab568d", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Gaming the Judge: Unfaithful Chain-of-Thought Can Undermine Agent Evaluation", "url": "https://arxiv.org/abs/2601.14691", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f48377990acdff9b651c9777097a4e1e615e75bf007047f90e059e03f37dcc00", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Generality-Training of a Classifier for Improved Calibration in Unseen Contexts", "url": "https://doi.org/10.1007/978-3-031-43424-2_23", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "58c033bca9690bfa126752950c06aafd175644a4086fbefde560f7cb335b55ad", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "GenRubric: Self-Evolving Rubric Generation for Scalable LLM Evaluation", "url": "https://arxiv.org/abs/2608.29856", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d53fde3c53380515d93354c3d7b6c8c712a928fa1aee6b9ce7d79995cdf5457c", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Geometry-Calibrated Conformal Abstention for Language Models", "url": "https://arxiv.org/abs/2604.27914", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "49d93a6bfa2aeecc99b61d9874d1b273e83afd83374fa9f92420cc6402f44db4", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "GMTRouter: Personalized LLM Router over Multi-turn User Interactions", "url": "https://arxiv.org/abs/2511.08590", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d15ded55ff4ed71d0f718ba9e9310b37391ec3ad0b68a6cb8e4bc7ae3ff870a1", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Governed Agentic Process Automation: A Floor-Safety Guarantee for Compliance-Critical LLM Routing", "url": "https://doi.org/10.3390/a19080627", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "019b8d248a89b5c21f2d65dbb839012e250c6be19cc01ff8acefa838a388cf44", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Grammar-Constrained Decoding Makes Large Language Models Better Logical Parsers", "url": "https://doi.org/10.18653/v1/2025.acl-industry.34", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "58c4be9c59ca85b42b5af4c947860c08480214578c70102f73002ffd45830ce4", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "GrammarRL: Effective Grammar-Constrained Decoding via Reinforcement Learning", "url": "https://arxiv.org/abs/2609.39869", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0af2bfae2a2c3eb14518f79b01fbbdfd0ea4611bac4cd8f45c0b0d239e9987b8", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "GraphPlanner: Graph Memory-Augmented Agentic Routing for Multi-Agent LLMs", "url": "https://arxiv.org/abs/2604.23626", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "842133cc09bf79bcd436600a3f4fc4921af6ab1995d34e2c1e05490f78946f05", "retrieved_at": "2026-10-16T10:01:21+00:00", "ledger": "research-sources.jsonl" }, { "title": "Hallucination as output-boundary misclassification: a composite abstention architecture for language models", "url": "https://arxiv.org/abs/2604.06195", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6de70a2182a04a49cabcb823c76638941a22de2c3f43058e15a71d6eeaf5cee8", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "Hallucination Detection and Confidence Calibration for Large Language Model Outputs: Reproducible Experiments on HaluEval", "url": "https://doi.org/10.69987/aimlr.2025.60401", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "03fd1f3cd24de7f2093947f6193564227ac4701667b53f4f07e74580fd431989", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "HAPS: Hierarchical LLM Routing with Joint Architecture and Parameter Search", "url": "https://arxiv.org/abs/2601.05903", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "41f7079ccdf9105481840b10e1383a8d1138efa436e3df36e22a1d1dfca6fd94", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routing", "url": "https://arxiv.org/abs/2405.00467", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "be2a3c079df22fb7211561b3fe514978131bd93d8cfeb58f1619b7b656d5d618", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Hierarchies of Calibration: Classification meets Regression", "url": "https://arxiv.org/abs/2606.03245", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6fa2f48847093eb41ddafde49c75f89786604ac2695e80a2af9256e23531e438", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answering", "url": "https://doi.org/10.1162/tacl_a_00407", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3fbb00ae0ff5e1ce42502dd2e6530165b50e5adf9ccdf6f32d0ee7f546edde6c", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "How Reliable Is Your Jailbreak Judge? Calibration and Adversarial Robustness of Automated ASR Scoring", "url": "https://arxiv.org/abs/2606.25487", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e653a3e82be985322baa95c2dcbed0c349bce6553cea397b667f68d2d0d281d0", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Human-Aligned Calibration for AI-Assisted Decision Making", "url": "https://arxiv.org/abs/2306.00074", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "578f85e6ec6fc8240122081390bb11a2ced35b22993b8db8bfe7b34d5737484a", "retrieved_at": "2026-10-16T10:02:42+00:00", "ledger": "research-sources.jsonl" }, { "title": "Human-Alignment and Calibration of Inference-Time Uncertainty in Large Language Models", "url": "https://arxiv.org/abs/2508.08204", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "caf025923569250c73ac49f52c6bc6bb8529ef907d5baed31245bce7f08a018c", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Human-Alignment, Calibration, and Activation Patterns in Large Language Model Uncertainty", "url": "https://arxiv.org/abs/2605.30675", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ead8997d6c847adcf349370893cbe89d361f7800404a2a2cb804f2cdb53c596d", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Improving Accuracy and Calibration of Deep Image Classifiers With Agreement-Driven Dynamic Ensemble", "url": "https://doi.org/10.1109/ojcs.2024.3519984", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4e5efb0d92f557d59f6f2063a5432e72fb51373fdfba902d461fcdfde8d7fc20", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Improving Multi-Class Calibration through Normalization-Aware Isotonic Techniques", "url": "https://arxiv.org/abs/2512.09054", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "9079c32b1036796832013dc775847dd6fb4c13b0722f22cfa2ff9c4f44494dde", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "In Defense of Softmax Parametrization for Calibrated and Consistent Learning to Defer", "url": "https://arxiv.org/abs/2311.01106", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "34bd6e47ad7fe905a7c547684deaa9b6b489d633e03b87c1c143f762417960f7", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Incentivizing Agentic Reasoning in LLM Judges via Tool-Integrated Reinforcement Learning", "url": "https://arxiv.org/abs/2510.23038", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a5c18b5cb4a4b688cbc26a7ca13cceec1391d2bcb3cdba01b1634d7a9bcd706e", "retrieved_at": "2026-10-16T10:01:17+00:00", "ledger": "research-sources.jsonl" }, { "title": "Independence and Calibration in Decision Analysis", "url": "https://doi.org/10.1287/mnsc.24.3.320", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "57146354c03b34b6fa13014de6bb26e06bb8bf27cb9aec499227b476f1be927c", "retrieved_at": "2026-10-16T10:02:42+00:00", "ledger": "research-sources.jsonl" }, { "title": "Inference Calibration of Neural Network Models", "url": "https://doi.org/10.1109/ickg55886.2022.00051", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "8ad1e736af067b3b83931403cab7e9e633e22133056a32fae278aa50a5c5d929", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "Intent-Description Anchoring Bias in LLM-as-a-Judge Evaluation of Recommender Systems", "url": "https://doi.org/10.1145/3773078.3841257", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "80be09f2af96a6f80f85c82a9891891822b081b382ab940197b63eb48c76764a", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Investigating Non-Transitivity in LLM-as-a-Judge", "url": "https://arxiv.org/abs/2502.14074", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "811d533ff382c8eb6aca484af3e6916315febc0070fa9a027e1694f038f614b8", "retrieved_at": "2026-10-16T10:01:15+00:00", "ledger": "research-sources.jsonl" }, { "title": "Investigating the Multilingual Calibration Effects of Language Model Instruction-Tuning", "url": "https://arxiv.org/abs/2601.01362", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "eb00622f0dee23410e19d62cf2ee6de58af2aa2bbbdcb68334e4eb88c4e1a4d8", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Investigating the Vulnerability of LLM-as-a-Judge Architectures to Prompt-Injection Attacks", "url": "https://arxiv.org/abs/2505.13348", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "04b6564ed5f7735db77511c50ef350671ef3d6b9416bd59ca4776c25367c1d41", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Investigating Uncertainty Calibration of Aligned Language Models under the Multiple-Choice Setting", "url": "https://arxiv.org/abs/2310.11732", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4b7afddf4cf0d9b6f01cacadc3c0938d5f52d5762f3f6e2692309aa31df9353e", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "iRULER: Intelligible Rubric-Based User-Defined LLM Evaluation for Revision", "url": "https://arxiv.org/abs/2602.12779", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f2788816acba4c1b1afe0792aa1bd31c5c90accdfe4099e366f59a6dfbff70ba", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Judging a Review by its Cover: A Reliability Analysis of LLM-based Peer Review Evaluation Metrics", "url": "https://arxiv.org/abs/2609.23264", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "8989bb42b1286d062ff74220aec83459595027fd929021a5812be7a018bff504", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "Judging the Judges: A Systematic Evaluation of Bias Mitigation Strategies in LLM-as-a-Judge Pipelines", "url": "https://arxiv.org/abs/2604.23178", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cb45bc018d91980cba2263acd8507f3265dd9359910a6f2631134037fc6f58d0", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Learned-Rule-Augmented Large Language Model Evaluators", "url": "https://arxiv.org/abs/2512.01958", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e7bafb0e336f2f7197212c63a03dcdabf4dd2189c0fe9903b979edb40560ea86", "retrieved_at": "2026-10-16T10:03:04+00:00", "ledger": "research-sources.jsonl" }, { "title": "Learning an Efficient Multi-Turn Dialogue Evaluator from Multiple Judges", "url": "https://arxiv.org/abs/2508.00454", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b74a44e0fa9d4bd4289155e2cb854d6846f4f68e1a845a63f34e67a0e5e4c699", "retrieved_at": "2026-10-16T10:01:17+00:00", "ledger": "research-sources.jsonl" }, { "title": "Learning Context-Free Grammars for Grammar-Constrained Decoding via Declarative Agentic Programming with Guarantees", "url": "https://arxiv.org/abs/2608.05493", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f731c3173eaebd7397e79873f6e38a9d736bbee1827164e08d3e7d5717f09af5", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "Learning to Defer in Congested Systems: The AI-Human Interplay", "url": "https://arxiv.org/abs/2402.12237", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d09158a0b1b46bf3b50f9762d32650b04ac9845411f4f2234be38444881de986", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Learning to Defer to Multiple Experts: Consistent Surrogate Losses, Confidence Calibration, and Conformal Ensembles", "url": "https://arxiv.org/abs/2210.16955", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c670606ea9dcf6db51a2ae4bd321d937a0be2ca6c230dd18868d2b9d0c107f43", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Learning to Defer with Better Features: A Case for Going Beyond the Final Layer", "url": "https://doi.org/10.23919/fruct67853.2025.11239207", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "98907a98dfdff61236d7403e5d00737139ab24c6a1add67fe872c1827d413c13", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Learning to Plan & Reason for Evaluation with Thinking-LLM-as-a-Judge", "url": "https://arxiv.org/abs/2501.18099", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "75f14754b1704903566d563a1ef0058062374034ace1c5e50d1e685ff7e46801", "retrieved_at": "2026-10-16T10:01:17+00:00", "ledger": "research-sources.jsonl" }, { "title": "Legibility is Not Interpretability: Comparing Judged and Actual Importance in Chain-Of-Thought Reasoning", "url": "https://arxiv.org/abs/2609.04194", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c0922f6d70dd80ee7e475c381ef41ca3068f2f2503ba6897547cd4cda93c7d25", "retrieved_at": "2026-10-16T10:01:21+00:00", "ledger": "research-sources.jsonl" }, { "title": "LH-Bench: Skill-Grounded Evaluation of Long-Horizon Agents on Subjective Enterprise Tasks", "url": "https://arxiv.org/abs/2603.22744", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "48e07a465175e29ba0e50d0aa0f4716919e122468a118cad6bb2c7896bb0d68d", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "LitCab: Lightweight Language Model Calibration over Short- and Long-form Responses", "url": "https://arxiv.org/abs/2310.19208", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "650fdcc06e034d1080c3a933b4a35381643174ec8a41a133ffae329fc18b1e8e", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "LLM Routing as Reasoning: A MaxSAT View", "url": "https://arxiv.org/abs/2603.13612", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3c7a1072da24ed6b5669b011a4bc7921debf261d272141139c966e97f23ae031", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "LLM-as-a-Judge for Human-AI Co-Creation: A Reliability-Aware Evaluation Framework for Coding", "url": "https://arxiv.org/abs/2604.27727", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "36d65493e5563765520466af0609fdbac8926e2c2aaea904712aac8c01f73d33", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "LLM-as-a-Judge for Scalable Test Coverage Evaluation: Accuracy, Operational Reliability, and Cost", "url": "https://arxiv.org/abs/2512.01232", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c49dcbb6ea1997265edbe70dad97b86e65bbade97f19c61c91ae732424736be9", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "LLMRANK: UNDERSTANDING LLM STRENGTHS FOR MODEL ROUTING", "url": "https://arxiv.org/abs/2510.01234", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "330f683ee8bd9d3e7277a62606ab2ea6feb84b9035bad231f21614b6e268d99e", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "LLMs Cannot Reliably Judge (Yet?): A Comprehensive Assessment on the Robustness of LLM-as-a-Judge", "url": "https://arxiv.org/abs/2506.09443", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "7065c256ddae877907034ad04688bfbcf2d92a6283efd6bd80db884e6196bdd6", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "LWCal: Loss-Weighted Calibration for Tabular Classifiers with Noisy Calibration Labels", "url": "https://arxiv.org/abs/2609.26839", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "22f77d6356edafde13e4c98bda219a9083a17510f6efb540447e047e4afc9e35", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Mediocrity is the key for LLM as a Judge Anchor Selection", "url": "https://arxiv.org/abs/2603.16848", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0fbbfbc3eca229a67e6152b11c7e17f74c59703883b4ad81ae6f846e98559606", "retrieved_at": "2026-10-16T10:01:17+00:00", "ledger": "research-sources.jsonl" }, { "title": "MESS+: Dynamically Learned Inference-Time LLM Routing in Model Zoos with Service Level Guarantees", "url": "https://arxiv.org/abs/2505.19947", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "662bc78dfb7ca3e9646cadedf210664bace09e50b351ad7020484347f5061a7c", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Mind the Confidence Gap: Overconfidence, Calibration, and Distractor Effects in Large Language Models", "url": "https://arxiv.org/abs/2502.11028", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "494a4679cf464ae1a25acca0d26c7ef38151c8ee2b2747e8a476abd24582750e", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "MIRROR: A Hierarchical Benchmark for Metacognitive Calibration in Large Language Models", "url": "https://arxiv.org/abs/2604.19809", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "736569e97f9dadfb7df0100c169b19078e1f4747638d2a347ab04a1368b8d4c6", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Mitigating Intrinsic Hallucinations Using Orthogonal Fine-Tuning and Dirichlet Prior Calibration", "url": "https://doi.org/10.1109/wccst67302.2026.11495718", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c6f1f9874578e889cff6daa10fc71cf661ced2e588e7904b43e473e3078f2682", "retrieved_at": "2026-10-16T10:02:28+00:00", "ledger": "research-sources.jsonl" }, { "title": "Mitigating Scoring Bias in LLM-as-a-Judge via Random Number Generation", "url": "https://arxiv.org/abs/2608.05726", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0a16099be8539131ab6f7437c5c7ea49120b170a22f25a93dfabd3d1a840298b", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Mitigating the Multiplicity Burden: The Role of Calibration in Reducing Predictive Multiplicity of Classifiers", "url": "https://arxiv.org/abs/2603.11750", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "7e7646892d8533bb54b518bc05cbed24bfd14876e7d03bd01920473922f29c2b", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Moment Multicalibration for Uncertainty Estimation", "url": "https://arxiv.org/abs/2008.08037", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "63d51befd7d276b6d73758d6e38f066b40ce67f61803b13ee53aae4c2eb2cdf2", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "MonoRouter: Enforcing Monotonicity in Item Response Theory for LLM Routing", "url": "https://doi.org/10.1109/gaiis69281.2026.11519177", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ebedcc344e7a2abeb6b45847543fcca3497e4bdfc561591e7437f02fbd480375", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Most of the LLM Routing Gap Is Task Type", "url": "https://arxiv.org/abs/2608.23023", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1349657e50d5d52d16a129ccaacb36c613b0855c4fd24f486a56e6dccc97df52", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "mR3: Multilingual Rubric-Agnostic Reward Reasoning Models", "url": "https://arxiv.org/abs/2510.01146", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2aee64207613b8bf83887eb887c6fbfd3ed35a9aab905b4b55ff2d0defb2c2ff", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Multi-Object Classification via Crowdsourcing With a Reject Option", "url": "https://arxiv.org/abs/1602.00575", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d65572ce39ab2a76512272beb59d41473b0a5aa72c7d99dbcca9bfedf1b9471a", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Multicalibration Boosting: Theory, Convergence, and Transferability", "url": "https://arxiv.org/abs/2605.24364", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "32f27c5cf5fb32ef3eea512dc671593c2d847acd5aef6f4fb3a58eeeb27b4328", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "Multicalibration for Confidence Scoring in LLMs", "url": "https://arxiv.org/abs/2404.04689", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5931b8192047b659f70c20ffa230354fce899e67f37fe9245f4843dc8a436424", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "Multicalibration for Unbiased Model-Based Prevalence Estimation", "url": "https://arxiv.org/abs/2604.21549", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e35b80b827f2352b6ae89b0dd9d3ee6e65ca08a9eed79b0cdc915f9cdb2bc1a7", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "Multicalibration in Fair Link Prediction", "url": "https://doi.org/10.1109/tcss.2025.3600263", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cf45be6217c7934b26da165b77da736dac3fff98b01f83da3ca2460d32932cca", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "Multiobjective Evolutionary Calibration of Uncertainty Quantification for Large Language Models", "url": "https://doi.org/10.1145/3795101.3805422", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "73499bbc89d7d1b8430f9f3a11a1585632e5838a5bdde5e4e38bf8d21f9fe0ce", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Multiprototype-based fuzzy classification and reject options", "url": "https://doi.org/10.1109/fuzzy.1996.552754", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "622d75f464c74a702ac28bd38ab150a9aba142d8fa8a979274159789bc377b26", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Naturally constrained reject option classification", "url": "https://doi.org/10.1007/s00138-024-01620-5", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6cd5bdee56cb53533f6bf09953dff645d804e8440806dfc5802642200f487e4c", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "No Need for Learning to Defer? A Training Free Deferral Framework to Multiple Experts through Conformal Prediction", "url": "https://arxiv.org/abs/2509.12573", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a46531cf335e77a06c3c69ff1306f8d7cd68922797cef6e05835f052334af81b", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Node Classification With Integrated Reject Option", "url": "https://arxiv.org/abs/2412.03190", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "75fc46e19d5e1c3caa770593f6f7b1c23674b393b82efed766fc41f0a948b18b", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "OdeBERT: One-stage Deep-supervised Early-exiting BERT for Fast Inference in User Intent Classification", "url": "https://doi.org/10.1145/3587464", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b263964c67e31c4c53e7c7775859faab5cc1b081a282adf1e478837e37992db0", "retrieved_at": "2026-10-16T10:03:10+00:00", "ledger": "research-sources.jsonl" }, { "title": "On Calibration of Large Language Models: From Response To Capability", "url": "https://arxiv.org/abs/2602.13540", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0d48cd98148d04b2dbd4e68c5556349ee23c9b87b32d28bddbf4e321bca34e3b", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "On Calibration of Mixup Training for Deep Neural Networks", "url": "https://arxiv.org/abs/2003.09946", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f5900280cd66d2a3efecdfba53b2867ecbcaf25609e0b4311a9fff0a266d5139", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks", "url": "https://arxiv.org/abs/1905.11001", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "9d7fc9694bb36af567d8dafbfb721c76a58b6d64e66c4cfe66057402d4fcfc09", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "On Reject and Refine Options in Multicategory Classification", "url": "https://doi.org/10.1080/01621459.2017.1282372", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b339e001f578ddadfff772ac1edc2141e5af45db9d6ab58dbfe695f2489333b0", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "On the Calibration of Large Language Models and Alignment", "url": "https://arxiv.org/abs/2311.13240", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "456a148737e13fea5462d0fa4b89d07d0f1c0ca6e29fa7b6a71bdeee2016f11f", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "On the Calibration of Massively Multilingual Language Models", "url": "https://arxiv.org/abs/2210.12265", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4730c324059f0fc904fb22e015acd6a0412be23774dffc7ed61723d330ac2dae", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "On the Calibration of Pre-trained Language Models using Mixup Guided by Area Under the Margin and Saliency", "url": "https://arxiv.org/abs/2203.07559", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "711f0cf7747ea4cd21aea37acfcf9ba0fc2401a758d5713fd07e31c51a1172a7", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "On the Dark Side of Calibration for Modern Neural Networks.", "url": "https://arxiv.org/abs/2106.09385", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "95b9cb45cbf4540ab0e2e5a20956f3dbe346f78d6032ba73c5998f77e520c236", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "On the Global and Local Calibration of Graph Neural Networks", "url": "https://vbn.aau.dk/da/publications/4eed889a-812e-4e3b-a276-23f7e5918232", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2f3952ab426d18ae30463e9257cd422dcc6135d795699d4e87759fac88af8e7d", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "On the Usefulness of the Fit-on-the-Test View on Evaluating Calibration of Classifiers", "url": "https://arxiv.org/abs/2203.08958", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3a0bfc84b612d51a1a6467a137ddc2e769e30a8e7952a0ec914066306c302807", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "On using focal loss for neural network calibration", "url": "https://ora.ox.ac.uk/objects/uuid:5fd484f2-b605-4256-8982-e5c10949c875", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1af15c2c11eb2b5bf4829fb2c55706d7518fcd34ccda51e17b95c739a43032fa", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "One Token to Fool LLM-as-a-Judge", "url": "https://arxiv.org/abs/2507.08794", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c13c11bfc88ef762389fd3ca345f889c22388a06e489690e2071f2b0e83a4a30", "retrieved_at": "2026-10-16T10:01:01+00:00", "ledger": "research-sources.jsonl" }, { "title": "Online Rubrics Elicitation from Pairwise Comparisons", "url": "https://arxiv.org/abs/2510.07284", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "997c504039e0b8d571509252da16678c361533365333e3a99b035b6e97fe3983", "retrieved_at": "2026-10-16T10:01:05+00:00", "ledger": "research-sources.jsonl" }, { "title": "OpenRubrics: Towards Scalable Synthetic Rubric Generation for Reward Modeling and LLM Alignment", "url": "https://arxiv.org/abs/2510.07743", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "273ab5baf9e29b7ef0d114a256efff94784af87df8079c71aae433d333999634", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Optimal Crowdsourced Classification with a Reject Option in the Presence of Spammers", "url": "https://arxiv.org/abs/1710.09901", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "70da9d4c522a83afe252408c5e5f2bd4d6e8743cba07c14551a6096a4d21b00d", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Optimal Design for Active Preference Learning with Biased LLM Judges", "url": "https://arxiv.org/abs/2609.38860", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cfda8c9b025517ff223bf675c15d35912795437245960d1be2649d82f6cd6e1f", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Optimal Deterministic Multicalibration and Omniprediction", "url": "https://arxiv.org/abs/2606.20557", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6229d7f0397c8c19a366530688d818b754ba3d32f710d9a8e698a0fcd79d4783", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "Optimization-based Prompt Injection Attack to LLM-as-a-Judge", "url": "https://arxiv.org/abs/2403.17710", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "bc0927fc7c1ff3b359cbbd29dcccf6883fe25126c7290cf2f771b601981724f3", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Optimizing Active Learning in Vision-Language Models via Parameter-Efficient Uncertainty Calibration", "url": "https://doi.org/10.1109/ijcnn64981.2025.11228515", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6f1c8b628e97c0b4f582ad7c1df8fe94654a937946fefe838e3a971ff4964e81", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Optimum Reject Options for Prototype-based Classification", "url": "https://arxiv.org/abs/1503.06549", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5aa3fadb879d3f8649d1a2e3d694b55b51e1e037cc4419295f6f7be11a35a449", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Play Favorites: A Statistical Method to Measure Self-Bias in LLM-as-a-Judge", "url": "https://arxiv.org/abs/2508.06709", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6553ef48a4528b1988d9ad31db35a5afd50f167f2ca8e23705d7f5ec5e69f1a9", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Post-hoc Calibration of Neural Networks.", "url": "https://arxiv.org/abs/2006.12807", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d212f8a5c8b1377420eca0ee9514f04297301b08f5046f8e1e306d6c9f231035", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "Posterior calibration and exploratory analysis for natural language processing models", "url": "https://arxiv.org/abs/1508.05154", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "be2d68645415fe5a1b8f01808094f08aee11db954ad26e98180b09d6335e87f5", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Power-controlled reliability tests for assessing probabilistic classifier calibration", "url": "https://doi.org/10.1080/03610918.2025.2574482", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cca27c5856d364b39008d05629afc08d65e4bb42d24e9d1b522cd89e7408f35c", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "PRD: Peer Rank and Discussion Improve Large Language Model based Evaluations", "url": "https://arxiv.org/abs/2307.02762", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "48c7fe48b77f97a3d3fa96b80d097638639572a01f1981aff4989db9c1d04ca2", "retrieved_at": "2026-10-16T10:01:04+00:00", "ledger": "research-sources.jsonl" }, { "title": "Predict Responsibly: Improving Fairness and Accuracy by Learning to Defer", "url": "https://arxiv.org/abs/1711.06664", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5a62851cae66b896b7956b3ef0f031b30431fb1f348b14552bc8f23b44d0fb4a", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Pricing Online LLM Services with Data-Calibrated Stackelberg Routing Game", "url": "https://arxiv.org/abs/2511.09062", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "9f5639de144b7ae165b397a4050db4d01312e6107dc6df54885b28e6f165f680", "retrieved_at": "2026-10-16T10:01:21+00:00", "ledger": "research-sources.jsonl" }, { "title": "Privacy Guard & Token Parsimony by Prompt and Context Handling and LLM Routing", "url": "https://arxiv.org/abs/2603.28972", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c531b5b07939e750df7290124eee7bc4033d01749bee3108483edf586f7315a6", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Privacy-Preserving LLMs Routing", "url": "https://arxiv.org/abs/2604.15728", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0e0cca7ff1c279a1b4b6f78e0497a6af1573b85bcc5ec97a759e63670a21f934", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Probabilistic Calibration by Design for Neural Network Regression", "url": "https://arxiv.org/abs/2403.11964", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "471f04fbfe199cc0798bcd03fd5e41a3096ac59e1a19a700d6355817d5a3b393", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "Probabilistic Calibration Is a Trainable Capability in Language Models", "url": "https://arxiv.org/abs/2605.11845", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "34618b252a0a0e90fd5b3f8d415e628029cc1cd46c884446bf8ed3075fe6904a", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Probabilistic Scores of Classifiers, Calibration is not Enough", "url": "https://arxiv.org/abs/2408.03421", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6e2725af221a74515f8e6601e305090eafbfddffc484f11a6d2befa5ca565bf9", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Proper-Composite Loss Functions in Arbitrary Dimensions", "url": "https://arxiv.org/abs/1902.06881", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b0575f916f3b49e1e44cf2f938a01401289f9f1a01d17488c43d2eb73c4a596e", "retrieved_at": "2026-10-16T10:02:39+00:00", "ledger": "research-sources.jsonl" }, { "title": "QA-Calibration of Language Model Confidence Scores", "url": "https://arxiv.org/abs/2410.06615", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a0133bc860af7a5f9c3e8e720ac95b15ac3b214ee551b88e5b3e75ca4e9810bf", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "QEDBENCH: Quantifying the Alignment Gap in Automated Evaluation of University-Level Mathematical Proofs", "url": "https://arxiv.org/abs/2602.20629", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b481e92d0f7dec2b4f18fa37c85c8ce309c70f561d92728de04dbbaebf9b2b5e", "retrieved_at": "2026-10-16T10:01:21+00:00", "ledger": "research-sources.jsonl" }, { "title": "Quality-of-Service Aware LLM Routing for Edge Computing With Multiple Experts", "url": "https://arxiv.org/abs/2508.00234", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e09f4b832585f91c309e4c5824c863e99af338033f109ace4bfb410229a1d956", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Quantifying and Mitigating Self-Preference Bias of LLM Judges", "url": "https://arxiv.org/abs/2604.22891", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "920f26e45143f93e66b8473ebe14b0d966feb83093515c7d9d5a787bdf77e5b0", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "RADAR: Reasoning-Ability and Difficulty-Aware Routing for Reasoning LLMs", "url": "https://arxiv.org/abs/2509.25426", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "da7fbb8bd32520aa20dd81fa67ebca3136fe6ee96d0195368b4b00444271e57b", "retrieved_at": "2026-10-16T10:01:25+00:00", "ledger": "research-sources.jsonl" }, { "title": "Realizable $H$-Consistent and Bayes-Consistent Loss Functions for Learning to Defer", "url": "https://arxiv.org/abs/2407.13732", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "28ba0c087ddf3bc97893be99c52840367003b5ff9e16abac0b35957e2e734c95", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Reasoning Model Is Superior LLM-Judge, Yet Suffers from Biases", "url": "https://doi.org/10.18653/v1/2026.evaleval-1.13", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "bda797091d1a87683cd19132da10c54d9f6c6b8beab4569b81a27ffac226b55b", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "ReCal: Reward Calibration for RL-based LLM Routing", "url": "https://arxiv.org/abs/2606.12479", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b05484357202a5f26476f7c32c131fee7d64d267751154fd6ce1bb9ab43a5f3b", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "RECAP: Relation Evidence Calibration for Detecting Spatial Relation Hallucinations in Vision-Language Models", "url": "https://arxiv.org/abs/2609.26093", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d49e957657854ae9d9e5e6acbcabdebaad500ed685c4ba5c2deeaa23ea83279d", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Reject option for VQ-based Bayesian classification", "url": "https://doi.org/10.1109/icpr.2000.906016", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ef7cbf9104af4395b6ad01d1d1dbb02de3eaca62857e7a3378a66760955cf36f", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Reliability without Validity: A Systematic, Large-Scale Evaluation of LLM-as-a-Judge Models Across Agreement, Consistency, and Bias", "url": "https://arxiv.org/abs/2606.19544", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2b70eaa63431891e5cff0ca5b796b2b6b460707e86539d4ee9743179a64a3ba0", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Reliable Multilabel Classification: Prediction with Partial Abstention", "url": "https://doi.org/10.1609/aaai.v34i04.5972", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c1fd0211c99b6ebc5d72181c201c4ce0b1ec33ca90d450382eaeb67c096e4f6b", "retrieved_at": "2026-10-16T10:02:44+00:00", "ledger": "research-sources.jsonl" }, { "title": "ReportLogic: Evaluating Logical Quality in Deep Research Reports", "url": "https://arxiv.org/abs/2602.18446", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "15ff5825903a423e55b601ad8f66f72e4537ad421a48fefc825949587b7eaba2", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "RerouteGuard: Understanding and Mitigating Adversarial Risks for LLM Routing", "url": "https://arxiv.org/abs/2601.21380", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cb76abff21afd7ffd6fca3c274e3d5a785f52359154a31be6ffb562407c5ac91", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Restoring Calibration for Aligned Large Language Models: A Calibration-Aware Fine-Tuning Approach", "url": "https://arxiv.org/abs/2505.01997", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ef311c36a030b34b3b169963549e14c1c29df8a69a9a0700ffeaef55f16549bc", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Retrieval Quality, Evidence, and Appropriate Abstention in Large Language Models", "url": "https://doi.org/10.64388/irev10i4-1723870", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f1e29aa3eeda59c63aadb60128f362d027a6753514443c64abf8ea0f1712ea54", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "Revisiting the Calibration of Modern Neural Networks", "url": "https://arxiv.org/abs/2106.07998", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "41aa64d05c9f3e88d947adeb0b1b94e463e2fb16fb103a17373b73b828620c48", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "Revisiting Uncertainty Estimation and Calibration of Large Language Models", "url": "https://arxiv.org/abs/2505.23854", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a992fa986e39469c8184ad10ceb9f293926b79ca4bf807fe87ff191952956734", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Reward-Based Online LLM Routing via NeuralUCB", "url": "https://arxiv.org/abs/2603.30035", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "13a5c92b1b0ecb3e2af1038137250ca12acb190bf154198a294640d7c99211b2", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "ROC-based cost-sensitive classification with a reject option", "url": "https://doi.org/10.1109/icpr.2016.7900146", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1f36b22234f3a4947e8603c3ca4d48ae931173242ce9c999b3d0db3932f24c5c", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "RouteNLP: Closed-Loop LLM Routing with Conformal Cascading and Distillation Co-Optimization", "url": "https://arxiv.org/abs/2604.23577", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f381a6b63c3ac44481f78589f9ed1626a44cb66d7715b946ce4ce67aa324bdfc", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Routing, Cascades, and User Choice for LLMs", "url": "https://arxiv.org/abs/2602.09902", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f04948f0630639c6c66e9b49d5cc8c4e3d519fe4051b80d020f23b3d1ce727fe", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "RSI-Router: Evolving Subtask-Level LLM Routing and Skills for Cost-Efficient Agents", "url": "https://arxiv.org/abs/2609.34712", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d1616c9398fc8324fb1c8bf344dd7b85448f67470bf2ae576d285327d502d598", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Rubric Is All You Need: Enhancing LLM-based Code Evaluation With Question-Specific Rubrics", "url": "https://arxiv.org/abs/2503.23989", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "27a481d2106d8f608452c3eaae73ec2bb3ecd64791f7ca9ac87724214e16b867", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "Rubric-Guided Evaluation Framework for Consistent Scoring with Large Language Models", "url": "https://doi.org/10.1109/iseee67817.2025.11304842", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f01e0bc8f740da0a9a8f279251a69e9677eabb8063efe584e687f0f0877697b3", "retrieved_at": "2026-10-16T10:02:57+00:00", "ledger": "research-sources.jsonl" }, { "title": "RubricBench: Aligning Model-Generated Rubrics with Human Standards", "url": "https://arxiv.org/abs/2603.01562", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c2748585a300bc3c6b6cf27207243b080d5743ffdcf9d73852f2700543d87ad7", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "RubricEval: A Rubric-Level Meta-Evaluation Benchmark for LLM Judges in Instruction Following", "url": "https://arxiv.org/abs/2603.25133", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "70dc40eee86a1266b5a12ae7aecdcdb12e75c30228cd9f3cab45b24e78645b6f", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "RubricRAG: Towards Interpretable and Reliable LLM Evaluation via Domain Knowledge Retrieval for Rubric Generation", "url": "https://arxiv.org/abs/2603.20882", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "c598e660f4f43ddef073da7da3282dc31312cd2134a7eb7c00ad289a59f53ac5", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Runtime Burden Allocation for Structured LLM Routing in Agentic Expert Systems: A Full-Factorial Cross-Backend Methodology", "url": "https://arxiv.org/abs/2604.01235", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5edc59ae3f6e345e201b4828e2a9ea387295713ed3842445480667dc3e3e4f5f", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Sample Complexity of Uniform Convergence for Multicalibration", "url": "https://arxiv.org/abs/2005.01757", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "648517e4da12ffe7a77124ee045e7110266c193b8764315a30d872afd900d032", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "Schema First Tool APIs for LLM Agents: A Controlled Study of Tool Misuse, Recovery, and Budgeted Performance", "url": "https://arxiv.org/abs/2603.13404", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "65a59078e507cdaabb1b0b83442eff89d3648ac2069e44f283669c8c3978d669", "retrieved_at": "2026-10-16T10:01:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "Second Guess: Detecting Uncertainty Through Abstention and Answer Stability in Small Language Models", "url": "https://arxiv.org/abs/2605.25394", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1c06843f82857728bc9a2ab37bdf667856f44bd19bc7649ce6da22861b82fbd2", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "SedarEval: Automated Evaluation using Self-Adaptive Rubrics", "url": "https://arxiv.org/abs/2501.15595", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2f39d719edd730749048a5f45c7d97a8402b475f96bf5900ebbcc82132c579fc", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Self- and Other-Labels Induce Bidirectional Bias in LLM Judges", "url": "https://arxiv.org/abs/2608.18091", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "694448391bee6de9814f38eee8bc70f6e0ed09628b3313e83452ef0ea99631f5", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Self-Preference Bias in Rubric-Based Evaluation of Large Language Models", "url": "https://arxiv.org/abs/2604.06996", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "aed66e65ea46a364025748e3e318791310d7cff021c3f3aea32a78acb7d862af", "retrieved_at": "2026-10-16T10:02:57+00:00", "ledger": "research-sources.jsonl" }, { "title": "Sensitive-Topic Leakage Through LLM Routing Metadata: Measurement and Mitigation", "url": "https://arxiv.org/abs/2610.09981", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cd52faf2faf0927069b4245becad3c035a4c848422e46e4d233cd605372ded9a", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "SeqRoute: Global Budget-Aware Sequential LLM Routing via Offline Reinforcement Learning", "url": "https://arxiv.org/abs/2605.25424", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f4e87aa929b6eaa82029e47f365147b4adc44d56cd65fa5251ded8c639bb77d0", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "SkewRoute: Training-Free LLM Routing for Knowledge Graph Retrieval-Augmented Generation via Score Skewness of Retrieved Context", "url": "https://arxiv.org/abs/2505.23841", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2ff86e47ffbe3f52bc2a0dc052188a08128e01c9f70bcf6745a12dc70539cd97", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Soft Calibration Objectives for Neural Networks", "url": "https://arxiv.org/abs/2108.00106", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "e6c88920834bd6073c6cd73443f3d86b9d6dbd933e5791f5b2ffeac1d6a7aa46", "retrieved_at": "2026-10-16T10:02:38+00:00", "ledger": "research-sources.jsonl" }, { "title": "STED and Consistency Scoring: A Framework for Evaluating LLM Structured Output Reliability", "url": "https://arxiv.org/abs/2512.23712", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a0b6c7549a4369b9f35f06c15e8daed190abdc2210e380f2b7922f69974b16c3", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "Stopping and Routing LLM Judge Panels", "url": "https://arxiv.org/abs/2608.19802", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "bdd0782a663b8ce9ed4a540e458eadd5e3fd51b7d3bec37d1631f28ccece62ed", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Structural Biases in LLM-as-a-Judge Systems: Implications for reliable IS Adoption", "url": "https://aisel.aisnet.org/amcis2026/sig_svs/svs/1", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "da82c3d7224eccce448d74196bfef43d1876ad604148c37bff0c91e89f7226fb", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Support Vector Rubrics: Closing the Gap Between Self-Generated and Human Rubrics", "url": "https://arxiv.org/abs/2606.08077", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "779ed6d428f21c17ba0f33d478b8ff29c0fe4e03c35f9991cc4bdcc8e58f2e61", "retrieved_at": "2026-10-16T10:01:19+00:00", "ledger": "research-sources.jsonl" }, { "title": "SwiftGram: Dynamic Lexical Forking for Precise and Efficient Grammar-Constrained Decoding", "url": "https://doi.org/10.1145/3832783.3834419", "category": "Research foundations", "topic": "structured_constrained_decoding", "description": "Comparison background for schema-constrained outputs and the distinction between structural validity and semantic correctness.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "4509acaaade503063f675bdccad31c32cc9c7e19597b30d3dcc98f03dc8e1136", "retrieved_at": "2026-10-16T10:03:11+00:00", "ledger": "research-sources.jsonl" }, { "title": "Taming Object Hallucinations with Verified Atomic Confidence Estimation", "url": "https://arxiv.org/abs/2511.09228", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "6e81bdaafccac11aa28dcc69ac1f38a4a5f57cbfdcdca7d7c4c57d2580325993", "retrieved_at": "2026-10-16T10:01:04+00:00", "ledger": "research-sources.jsonl" }, { "title": "Target-Aware Calibration Data Selection for Preserving Uncertainty in Quantized Language Models", "url": "https://arxiv.org/abs/2608.21019", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "40af9c9312208194ba7d988ba1d6d80fce5c44e7950683ab083337cf04de5f3f", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Task Abstention for Large Language Models in Code Generation", "url": "https://arxiv.org/abs/2605.17029", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3f76354d89cbdeaf2245f85516585f90ffe2fe9a0de7a6275c4e42895017d5c1", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "Task-Aware LLM Routing with Multi-Level Task-Profile-Guided Data Synthesis for Cold-Start Scenarios", "url": "https://arxiv.org/abs/2604.09377", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "09095bc6fcac312734627e05d980da4d7c5f3a52f15e223eb45c0e42dcea852b", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "The ACUTE Protocol: Operationalizing Language Model Activations for Better Calibration, Utility, and Trust", "url": "https://arxiv.org/abs/2606.07822", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a22c2a163038f1dc07bbe97dff34fe5c59eb1ad794698d8682d3659ea270cf51", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation", "url": "https://arxiv.org/abs/2606.13685", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cc5fdcd9d30a18c4fb41d9fb42b0b2bd0b26b5e4433d6fa0a2291715d3a2d6b2", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "The data replication method for the classification with reject option", "url": "https://doi.org/10.3233/aic-130566", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "72ca78f16bff9f456f7472d01d45da0fa2da0e292a8bd667987b1565664c862e", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "The illusion of a perfect metric: Why evaluating AI's words is harder than it looks", "url": "https://arxiv.org/abs/2508.13816", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "f8f168fff6ea0807babb7065959e32e0179c58b3c58dd780b02509dc613e19b4", "retrieved_at": "2026-10-16T10:01:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "The Matthews Correlation Coefficient (MCC) is More Informative Than Cohen’s Kappa and Brier Score in Binary Classification Assessment", "url": "https://doi.org/10.1109/access.2021.3084050", "category": "Research foundations", "topic": "proper_scoring_rules", "description": "Background for evaluating probability distributions and incentives for truthful probability reporting.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d7e7a30d9a7d9a0b96864d06671138d7bc0bf7492a843bef503a6e4385f2c937", "retrieved_at": "2026-10-16T10:02:41+00:00", "ledger": "research-sources.jsonl" }, { "title": "The Nearest Neighbor Classification Rule with a Reject Option", "url": "https://doi.org/10.1109/tssc.1970.300339", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "11909882ab808a71291a798e49d7cdee6c101714d7905f5a45a2364c2183cdb5", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "The Provenance Paradox in Multi-Agent LLM Routing: Delegation Contracts and Attested Identity in LDP", "url": "https://arxiv.org/abs/2603.18043", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3edfd41d53ebfc1395324af6397ffab4b5e3409a69a74111dce98ce1ebc7ebdb", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "The Reliability Paradox: Exploring How Shortcut Learning Undermines Language Model Calibration", "url": "https://arxiv.org/abs/2412.15269", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2d0a5ce6f3efc4e67e67180ee3b0887b5a027e0a5d7f83736879007138d6d281", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "The Silent Judge: Unacknowledged Shortcut Bias in LLM-as-a-Judge", "url": "https://arxiv.org/abs/2509.26072", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5fe0a4b05515664b07f1a7a464b772568aa20174864c47b2347b59fd3ab6ea2f", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Theory and Algorithms for Learning with Multi-Class Abstention and Multi-Expert Deferral", "url": "https://arxiv.org/abs/2512.22886", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "1a7d6cf8413e248fb08ee26f996f4a68cf0b7a99dd2a2a5d3ac4f498376a591e", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "To Steer or Not to Steer? Mechanistic Error Reduction with Abstention for Language Models", "url": "https://arxiv.org/abs/2510.13290", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "28dd19cb4fe6513ee87a5c775ed4af31e9e49774b0298e829a1a46c9652df4dd", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "Too Much of the Same: From Algorithmic to Human Bias in Learning to Defer", "url": "https://arxiv.org/abs/2608.28050", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b87799e27ed10165c096c25d8947f17409d7c578bc5acff5dae14d167f9c7908", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Towards Anytime Classification in Early-Exit Architectures by Enforcing Conditional Monotonicity", "url": "https://arxiv.org/abs/2306.02652", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b81c7b014502a5d3a689c38506f805bd87be80a35107ca2647635fdfcfe1115e", "retrieved_at": "2026-10-16T10:03:10+00:00", "ledger": "research-sources.jsonl" }, { "title": "Towards Calibrated Losses for Adversarial Robust Reject Option Classification", "url": "https://arxiv.org/abs/2410.10736", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d3efd93620373e5bc22a2995260b86db954e19bee202bba00ad88c418f76a976", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Towards Cost-effective LLMs Routing with Batch Prompting", "url": "https://arxiv.org/abs/2605.28268", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "2764ef799dfe1c6fa7b78811f940539ef7fbd637c1273a056596e6dcf984b7d9", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Towards Optimizing SQL Generation via LLM Routing", "url": "https://arxiv.org/abs/2411.04319", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a25d356aa0bf7209ccab91994e531bb28b153e106d50c2c6ac016addaadee4a4", "retrieved_at": "2026-10-16T10:01:22+00:00", "ledger": "research-sources.jsonl" }, { "title": "Towards Provably Unbiased LLM Judges via Bias-Bounded Evaluation", "url": "https://arxiv.org/abs/2603.05485", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "86ebcb8b54830af4a9f62d358d9b075491553ca58e04866025ff4ae06884f370", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Training LLM Judges from Language Feedback via Position-Selective Self-Distillation", "url": "https://arxiv.org/abs/2609.38792", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "90d3940743cefab43657448b6f570cdd7ee40354589a39d907312ad6870dfe08", "retrieved_at": "2026-10-16T10:01:13+00:00", "ledger": "research-sources.jsonl" }, { "title": "Two Ways to De-Bias an LLM-as-a-Judge: A Continuous-Score Comparison of Hierarchical Bayesian Calibration and Neural-ODE Score Transport", "url": "https://arxiv.org/abs/2605.09227", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "54e99fb0ef41751abb101780313e04d41fda168c46429d172e30424fc73c02f3", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Uncertainty in Language Models: Assessment through Rank-Calibration", "url": "https://arxiv.org/abs/2404.03163", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "cfa32ed1ee19e0707690a37598bdbb6fb1255b3d92713c433a80a0d5544451f9", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Uncertainty-Aware Abstention in Large Language Models with Provable Alignment Guarantees", "url": "https://arxiv.org/abs/2607.04430", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a94ed25ec3f7d1d9b6de16044c6c8dd6b249a96f236b45b78d669737fd5e8589", "retrieved_at": "2026-10-16T10:02:48+00:00", "ledger": "research-sources.jsonl" }, { "title": "Unlabeled Target-Domain Calibration for Tabular Classifiers under Label Shift", "url": "https://doi.org/10.1109/icassp55912.2026.11464109", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "50df31370026d21d5b040034b8e27666737a0ca2f24b3142ed0aa724fee26162", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "Unsolvability Ceiling in Multi-LLM Routing: An Empirical Study of Evaluation Artifacts", "url": "https://arxiv.org/abs/2605.07395", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "7f42c7ae335a250e0ccc32186f93ba1ef389893bb84d767440c7766da4538d8b", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "Unveiling Uncertainty: A Deep Dive into Calibration and Performance of Multimodal Large Language Models", "url": "https://arxiv.org/abs/2412.14660", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d4c5569a6c9e1f80a344d34ec86acc5c77d633763beaa7bb09c7c8feb67de953", "retrieved_at": "2026-10-16T10:02:36+00:00", "ledger": "research-sources.jsonl" }, { "title": "Using Natural Language Explanations to Rescale Human Judgments", "url": "https://arxiv.org/abs/2305.14770", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "aeca4e89b0d311f81824cfef7c0dedb817f9ce53d5b5315df9863e877a1c7df0", "retrieved_at": "2026-10-16T10:01:05+00:00", "ledger": "research-sources.jsonl" }, { "title": "Variable-Based Calibration for Machine Learning Classifiers", "url": "https://arxiv.org/abs/2209.15154", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d526d4eaad797763accedea3a44527ab52466106744ce0598836ca358f7bfca7", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "VDAR-Router: Adaptive LLMs Routing via Verbalized Query Difficulty Analysis Retrieval", "url": "https://arxiv.org/abs/2607.18098", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "60174752c8db21fb92485def00e02f2af285f7d7ce4895de314463f7bab5cbec", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "VeriRouter: Towards Verifiable LLM Routing", "url": "https://arxiv.org/abs/2610.17431", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "98cf07ca4d39a71d6819aa17a7e3759a8de3ef8038e6d4f4a641ed95edd51638", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" }, { "title": "VideoJudge: Bootstrapping Enables Scalable Supervision of MLLM-as-a-Judge for Video Understanding", "url": "https://arxiv.org/abs/2509.21451", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "5a4a443136e894b106a04cec43b59715499da91081e020d7b0cca41b0e123137", "retrieved_at": "2026-10-16T10:01:16+00:00", "ledger": "research-sources.jsonl" }, { "title": "What is Your Metric Telling You? Evaluating Classifier Calibration under Context-Specific Definitions of Reliability", "url": "https://arxiv.org/abs/2205.11454", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "724e62062d470b018998edccb8a114d04cd880f2c6bf09204bd105e8b241dc94", "retrieved_at": "2026-10-16T10:02:37+00:00", "ledger": "research-sources.jsonl" }, { "title": "When Consistency Does Not Mean Reliability: Evaluating Local LLM Judges Against Human Ratings", "url": "https://arxiv.org/abs/2609.13824", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "0bfcdf3eb4966b675d289b7222c9944da0d942e0c9cc1ad66fa5b8ff21e5562d", "retrieved_at": "2026-10-16T10:02:58+00:00", "ledger": "research-sources.jsonl" }, { "title": "When is Multicalibration Post-Processing Necessary?", "url": "https://arxiv.org/abs/2406.06487", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "3b362b9a6b40a226f21e4a6bcda0f61a900be5e360232196a39cce87718e7762", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "When More Experts Hurt: Underfitting in Multi-Expert Learning to Defer", "url": "https://arxiv.org/abs/2602.17144", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "ec52b7ff5e1f612382fa61045500cf71c5c6fae55f1e8c0856325fb54aa64eb5", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "When Noise Fabricates Bias: The Fragility of LLM-as-a-Judge Bias Measurement under Noisy Text", "url": "https://arxiv.org/abs/2609.11067", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "d6f12f151f51c34314628e8b8c2b03f2bb01b7f274a397a006425c7551b6305a", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "When Should LLMs Abstain? Chain-of-Self-Questioning for Selective Risk Control", "url": "https://arxiv.org/abs/2609.17516", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "89ac93951d8b0f7167755ca566d7572724ae4452bbddbc38dcbab0519485c18a", "retrieved_at": "2026-10-16T10:02:55+00:00", "ledger": "research-sources.jsonl" }, { "title": "When Wording Steers the Evaluation: Framing Bias in LLM judges", "url": "https://arxiv.org/abs/2601.13537", "category": "Research foundations", "topic": "judge_reliability", "description": "Background for evaluating the reliability, bias, consistency, or uncertainty of model-based judgments.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "a5058c23cfa7f8cf42e65fefe3df85ce29c5b8f36bd684498333ff591319908a", "retrieved_at": "2026-10-16T10:02:52+00:00", "ledger": "research-sources.jsonl" }, { "title": "Who Should Predict? Exact Algorithms For Learning to Defer to Humans", "url": "https://arxiv.org/abs/2301.06197", "category": "Research foundations", "topic": "selective_prediction", "description": "Background for confidence thresholds, abstention, deferral, and the risk versus coverage trade-off.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b639c6cdbe7a3a55a37ba5b528b0c9f9b9e97fac9cfa0ec972e3624934e663eb", "retrieved_at": "2026-10-16T10:02:46+00:00", "ledger": "research-sources.jsonl" }, { "title": "Who’s the (Multi-)Fairest of Them All: Rethinking Interpolation-Based Data Augmentation Through the Lens of Multicalibration", "url": "https://arxiv.org/abs/2412.10575", "category": "Research foundations", "topic": "probability_calibration", "description": "Background for checking whether reported decision probabilities agree with observed outcomes.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "b355006e4f04568d7e7ab3adbe90639adcaf688233f235e2c7e746defa1470ef", "retrieved_at": "2026-10-16T10:02:27+00:00", "ledger": "research-sources.jsonl" }, { "title": "WISERouter: LLM Routing with Workload Budget Constraint", "url": "https://arxiv.org/abs/2607.23765", "category": "Research foundations", "topic": "efficient_routing_cascades", "description": "Comparison background for allocating prediction work across classifiers or models under accuracy and compute constraints.", "relation": "background", "verification_level": "abstract_checked", "content_sha256": "807b7e071f1d793324b6b136074b75b453f84a44fe08d4441a710334da6bc4f7", "retrieved_at": "2026-10-16T10:03:07+00:00", "ledger": "research-sources.jsonl" } ], "excluded_from_target": [] }