Entity · paper

Looping Is Not Reliability: State-Bound Evidence and Typed Revision Contracts for Agentic Code Repair

paperactiveprovisionallooping-is-not-reliability-state-bound-evidence-and-typed-revision-contracts-for-agentic-code-repair-abf1ca5f·1 events·first seen 4d ago

Aliases: Looping Is Not Reliability: State-Bound Evidence and Typed Revision Contracts for Agentic Code Repair

Co-occurring entities

HumanEval

More like this (12)

Recursive Self-Improvement in AI: From Bounded Self-Refinement to Autonomous Research Loops Copy Less, Ground More: Overcoming Repetitive Copying in Long-Context Reasoning via Evidence-Aware Reinforcement Learning Operadic consistency: a label-free signal for compositional reasoning failures in LLMs Win by Silence: Deletion Non-Monotonicity, Autonomous Exploitation, and Typed-State Gating in LLM Plan Evaluation Valid Inference with Synthetic Data via Task Exchangeability Do Coding Agents Deceive Us? Detecting and Preventing Cheating via Capped Evaluation with Randomized Tests Can We Break LLMs Out of Self-Loops? Fine-Grained Reasoning Control with Activation Steering Form, Not Content? A Preregistered, Placebo-Controlled Evaluation of Learned Error-Conditioned Self-Repair Through Prompts and Weights in Frozen Small Code Models The Signal-Coverage Matrix: Stratifying Type and Semantic Errors in Statement Autoformalization Partially Correlated Verifier Cascades in LLM Harnesses: Concave Log-Odds, Polynomial Reliability, and Blind-Spot Ceilings verification-refinement loop GitOfThoughts: Version-Controlled Reasoning and Agent Memory You Can Replay, Diff, and Merge

Recent events (1)

5arXiv · cs.CL·4d ago·source ↗

Typed loop contracts for agentic code repair: evidence that stale traces degrade correctness retention

A new arXiv preprint studies the reliability gap in generate-test-revise loops used by coding agents, finding that forced revision cycles cause current correctness to drop from 0.820 to 0.673 even as ever-correct rises to 0.847. Controlled experiments with 2,430 branches show stale traces harm 34/135 correct starts versus 4/135 with current traces, a statistically significant 22.2-point increase. The authors formalize the problem by separating admission, preservation, and certification concerns, then derive a typed loop contract with a mechanically enforceable reference implementation that binds verifier evidence to exact code states. The work is framed explicitly as a specification artifact rather than a claim of improved repair competence.

Evaluation and Benchmarking AI Safety Research HumanEval Looping Is Not Reliability: State-Bound Evidence and Typed Revision Contracts for Agentic Code Repair +1 more