# REWIND: Uncertainty-Guided Trajectory Repair for Test-Time Reasoning

arXiv: https://arxiv.org/abs/2610.19561 · HTML: https://mlsanity.com/paper/2610.19561/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 14 of its 25 references that are in the map.
- First posted on arXiv 14 Oct 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 14 of 14)

- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [GPQA: A Graduate-Level Google-Proof Q&A Benchmark](https://mlsanity.com/paper/2311.12022/) (2023)
- [Language Models (Mostly) Know What They Know](https://mlsanity.com/paper/2207.05221/) (2022)
- [Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them](https://mlsanity.com/paper/2210.09261/) (2022)
- [Large Language Models Cannot Self-Correct Reasoning Yet](https://mlsanity.com/paper/2310.01798/) (2023)
- [An Explanation of In-context Learning as Implicit Bayesian Inference](https://mlsanity.com/paper/2111.02080/) (2021)
- [Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Models](https://mlsanity.com/paper/2310.04406/) (2023)
- [Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning](https://mlsanity.com/paper/2506.01939/) (2025)
- [Faith and Fate: Limits of Transformers on Compositionality](https://mlsanity.com/paper/2305.18654/) (2023)
- [Let's Sample Step by Step: Adaptive-Consistency for Efficient Reasoning and Coding with LLMs](https://mlsanity.com/paper/2305.11860/) (2023)
- [CGMH: Constrained Sentence Generation by Metropolis-Hastings Sampling](https://mlsanity.com/paper/1811.10996/) (2018)
- [Forking Paths in Neural Text Generation](https://mlsanity.com/paper/2412.07961/) (2024)
- [Entropy-based Exploration Conduction for Multi-step Reasoning](https://mlsanity.com/paper/2503.15848/) (2025)
- [A Block Metropolis-Hastings Sampler for Controllable Energy-based Text Generation](https://mlsanity.com/paper/2312.04510/) (2023)

## Abstract (arXiv)

An early mistake in the generated reasoning can constrain all subsequent reasoning. Repeated-generation methods, like MCMC-based approaches, attempt to improve output quality by sampling multiple generations, but their effectiveness can be limited by randomly selecting where to regenerate. On MATH500 with Qwen2.5-Math-7B, we find that incorrect responses have substantially higher mean token entropy than correct responses, with an average within-question relative increase of approximately 112%, suggesting that entropy can signal reasoning errors. Motivated by this finding, we introduce REWIND, a training-free controller that uses entropy dynamics to propose targeted suffix regenerations. Across three model families and five benchmarks, REWIND outperforms all baselines, with gains of 0.51-10.00 percentage points over the strongest baseline. These results demonstrate the effectiveness of targeted, uncertainty-guided trajectory repair across models and tasks. Our code is available at https://anonymous.4open.science/r/rewind-D310.