arXiv AI September 11, 2026 By Eshwar Reddy M, Sourav Karmakar

Proof-Carrying Cognition: Closing the Verification Gap with Reality-Settled Reward

The paper introduces the concept of proof‑carrying cognition, aiming to close the verification gap in language‑model reasoning by using reality‑settled rewards. It presents a theoretical framework linking verifier‑gold correlation to compute‑capability trade‑offs, demonstrates that unsound verifiers degrade under best‑of‑N selection while sound verifiers improve, and proposes a new benchmark metric, Soundness‑under‑Pressure, for evaluating reality‑settled reasoning systems.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

arXiv Machine Learning

Jun 19

Granularity-Regulated Adaptive Computational Efficiency for Optimal Verification in Test-Time Scaling

arXiv:2606. 19354v1 Announce Type: cross Abstract: Test-time scaling (TTS) has emerged as a powerful paradigm for improving the reasoning performance of large language models (LLMs) by investing additional compute at inference time.

By Ardit Krasniqi, Luan Vejsiu, Elira Dervishi

llms benchmarks

arXiv AI

Sep 1

Drop the Act: Probe-Filtered RL for Faithful Chain-of-Thought Reasoning

arXiv:2605.11467v2 Announce Type: replace-cross Abstract: Reasoning models post-hoc rationalize answers they have already committed to internally, producing chains of *reasoning theater*: deliberativ...

By Swapnil Parekh, Naman Goyal

llms reinforcement-learning benchmarks safety

arXiv AI

Aug 18

Measuring Reward Hacking and Reasoning-Answer Decoupling Under Position-Confounded Optimization

arXiv:2608. 15445v1 Announce Type: new Abstract: When a reward is correct on every training example yet consistent with more than one goal, a model can acquire an unintended one, a failure known as goal misgeneralization.

By Suyash Maniyar, Armaan Sandhu, Abhishek Mishra

llms benchmarks

arXiv AI

Jun 18

DeFAb: A Verifiable Benchmark for Defeasible Abduction in Foundation Models

arXiv:2606. 18557v1 Announce Type: new Abstract: A rule-based logic solver resolves every instance in our benchmark in under 50 microseconds with 100% accuracy; the best frontier language model reaches 65% at best and drops to 23.

By Patrick Cooper, Alvaro Velasquez

llms reinforcement-learning benchmarks

arXiv Machine Learning

Jul 8

More Convincing, Not More Correct: Self-Play Reward Hacking of Reference-Free LLM Judges

arXiv:2607. 05904v1 Announce Type: new Abstract: Training a language model against its own reference-free judgments (the premise of self-rewarding, self-play, and LLM-as-a-judge pipelines) assumes a model's verdict on a shown answer tracks correctness.

By Chenyu Zhou

llms reinforcement-learning

arXiv AI

Jul 14

LLMs as a Jury: Cross-Model Consensus Can Outperform Process Reward Models for LLM Reasoning

arXiv:2607. 10139v1 Announce Type: cross Abstract: Selecting the correct answer from a pool of candidate reasoning chains is the engine of test-time scaling, yet the standard selectors each carry a cost: self-consistency inherits the errors of the single model it resamples, and trained reward models need labeled data and transfer poorly off-distribution.

By Ning Liu

llms benchmarks

One email a morning, machine-written

Related stories

Granularity-Regulated Adaptive Computational Efficiency for Optimal Verification in Test-Time Scaling

Drop the Act: Probe-Filtered RL for Faithful Chain-of-Thought Reasoning

Measuring Reward Hacking and Reasoning-Answer Decoupling Under Position-Confounded Optimization

DeFAb: A Verifiable Benchmark for Defeasible Abduction in Foundation Models

More Convincing, Not More Correct: Self-Play Reward Hacking of Reference-Free LLM Judges

LLMs as a Jury: Cross-Model Consensus Can Outperform Process Reward Models for LLM Reasoning