Jev and typed decision model papers

Studies fitted confidence thresholds for sharing agent control with decision models.

A benchmark spanning matched requests and closed-loop tasks. Experiments have not been reproduced here.

In ten prepared scientific cases, the authors separate semantic selections, computed quantities and final labels; comparison-model errors can be hidden by unchanged final labels.

Gates reasoning start, candidate-answer reading and generation stopping; evidence is limited to a frozen Qwen backbone and three task families.

Tests classification and verification under content-preserving table rewrites across three task families; confidence can flag instability without proving correctness.

Audits ordinal label underuse across selected decision models; finer scales increase compression, while accuracy, final-label utilization and averaged candidate probabilities remain distinct measures.

Text and vision attacks with label-validity checks and held-out evaluation models.

A decision-model architecture exploring null handling, option permutations, scaling, and caching. Its design is Jeva’s, not Jev’s disclosed internals.

Structured video decisions with temporal spans and candidate scores; cached and joint inference reuse visual computation.

Pairs factual probes with action choices to diagnose decision failures.

Tool-argument judges including Jev reject some correct derived values without visible references; model-generated references are evaluated on three benchmarks.

A controlled adversarial comparison of Jev, LLM judges, and purpose-built safety classifiers.

Compares frozen Qwen decision readouts with optional KL-anchored fine-tuning and budgeted reasoning; reported gains depend on backbone size and routing workload, not disclosure of hosted Jev internals.

Compares rubric judges and examines correlated errors that can limit cascade improvements.

Answer-letter readout and supervised LoRA baselines; evaluation includes model-generated labels.

Bilingual decision evaluation and reliability when local decisions compose into larger systems.

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.

Looped hidden-state computation for typed decisions; additional loops trade compute for accuracy.

Candidate-menu interventions and confidence cascades; ranking stability does not establish probability calibration.

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.

ContractNLI request variations can change which answers are wrong while aggregate accuracy stays unchanged; stable answers may preserve errors. This benchmark does not validate real legal deployment.

Studies confidence-based escalation from Jev to a reasoning judge, including limits under distribution shift.

Tests option-channel attacks on seven open-weight models; reversed decisions can remain confident. The reported experiments do not test hosted Jev.

Knowledge, reasoning and multilingual multiple-choice evaluation of Jev; reported rankings and mathematical weaknesses apply to the selected models and benchmarks.

Bi-encoder approach to multimodal decisions and retrieval through natural-language interfaces.

Optimizes context additions around unchanged questions and options in four decision systems; results concern initially correct decisions, and human validity checks cover a sampled subset of successful attacks.

Tests conflicts between option names and definitions despite structurally valid outputs.

Same-number citation swaps probe financial-evidence verification; rejecting wrong roles trades off with accepting equivalent support in a constructed test.

Studies accuracy, cost, and calibration in seven political-science replication tasks.

A common-request harness compares typed decision models, generative readouts and trained classifiers; accuracy, selective coverage and out-of-scope acceptance depend on label access and task.

English/Arabic persona-conditioned Jev survey audit; the effects concern simulated respondents in one model, not measurements of real populations.

Examines reliability, group-specific confident errors, and selective automation in security decisions.

Turkish typed-decision evaluation reports decision quality, Brier-based probability quality and risk–coverage alongside input probes; most labels are AI-derived, and the lab’s model used feedback from earlier test results.

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.

Training-free decision readout with label-prior and option-order correction.

Composes finite action fields and explores alternatives using interaction feedback.

Shared evidence encoding and candidate-probability readout, trained with representation and action-outcome feedback.

An early benchmark study of Jev. Read the task selection and evaluation setup alongside its reported results.

A Chinese typed-decision model and proposed benchmark; domain-specific training calls for independent validation.

Keeps raw memory records, separating Jev relevance judgments from LLM search planning and answer generation.

A controlled calibration study using constructed probabilities. Interpretations of latent uncertainty remain hypotheses.

Compares Jev and generative judges on retrospective agent-trace risk classification; dataset rankings vary, and successful-call latency and valid-judgment cost must be read with coverage.

Tests coherence between logically related probabilities, separately from accuracy and calibration.

An evidence audit and a checklist for evaluating typed decision models.

A black-box adversarial benchmark. A changed decision is not necessarily an incorrect decision.

Estimates whether asking for missing information would improve a typed decision.

Benchmarks Jev alignment-failure detection while separating question wording from input context; benchmark-scorer labels dominate, and context gains often depend on fields encoding the label.

A Jev-style visual-choice study separating adaptation gains, candidate-readout validity, and calibration limits.

Audits Jev injection on reconstructed InjecAgent cases, separating probability shifts from attacker-target selections; optimization results are checked on fresh calls, within this benchmark setting.

Jev checks support between generated and reference radiology reports in both directions; expert agreement depends on report size and error definitions and does not establish prospective clinical effectiveness.

Jev state tables and sparse flagship-model anchors support population simulations with shared human–model discrepancy estimates; larger simulations do not add human evidence or establish prospective intervention effects.

Evaluates zero-shot social-science annotation and confidence-routed LLM cascades; lower costs coexist with task-dependent errors and overconfidence, so confidence cannot substitute for label validation.

Confidence-gated Jev control with LLM fallback; gains depend on action alternatives and baseline routing.

An independent visual decision architecture shares image context across batched questions using an LM-head readout; warm amortized speedups trade higher peak memory, while accuracy gains concentrate on trained task families.

Studies Jev as a controller for agent memory. Reported outcomes concern the complete memory workflow.

An open typed-decision model using hidden-state readout. The abstract reports a small external comparison.