Review Rarely Beats the Stronger Model: Separating Verification from Capability Transfer in LLM Pipelines
cs.CL, cs.AI
Submitted: 2026-10-11
Updated: 2026-10-14
Code: https://github.com/huggingface/math-verify
Terminology
Sources
- Seed-Coder: Let the Code Model Curate Data for Itself
- Commit-first LLM judging inherits the judge's own errors
- The Llama 3 Herd of Models
- Qwen2.5-Coder Technical Report
- JEV-as-a-Judge: Accept When Confident, Escalate When Unsure
- When to Screen, When to Bypass: LLM-Judges in Resource-Scarce AI-Human Workflow
- When Does Verification Pay Off? A Closer Look at LLMs as Solution Verifiers
- Easier to Mislead Than to Correct: Harmful and Beneficial Revision in LLM Conformity
- Cross-Model LLM Code Review: Should you use Claude to review Codex or vice versa?
- Qwen3 Technical Report
- Qwen2.5 Technical Report
- Precise but Uncoupled: Reviewer Precision Does Not Guarantee Critique Uptake in Multi-Agent Math Reasoning
- More Convincing, Not More Correct: Self-Play Reward Hacking of Reference-Free LLM Judges
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering