通过分步纠错提升大模型数学推理能力，准确率超主流方法2.4个百分点。
Enhancing Mathematical Reasoning in LLMs by Stepwise Correction
- 引入分步验证与修正机制，让模型自我发现并修改错误推理步骤。
- 在8个数据集上平均准确率达94.1%，较最优基线提升2.4个百分点。
- 减少生成路径数量，令牌消耗降低77.8%，更高效实用。
Best-of-N解码方法通过生成多个解题路径、评分并选取最高分结果来提升数学推理性能，但重复独立生成常导致相同错误，最终答案仍可能错误。本文提出新型提示方法Stepwise Correction（StepCo），使大模型能够识别并修正自身推理过程中的错误步骤。该方法迭代执行验证与修正阶段，采用过程监督验证器进行判断。验证-修正流程不仅显著提升答案正确率，还大幅减少所需生成路径数，降低令牌消耗。使用GPT-4o作为后端模型时，StepCo在八个数据集上平均准确率达94.1%，相比当前最佳的Best-of-N方法提升2.4个百分点，同时减少77.8%的令牌开销。
原文摘要 · Abstract (English)
Best-of-N decoding methods instruct large language models (LLMs) to generate multiple solutions, score each using a scoring function, and select the highest scored as the final answer to mathematical reasoning problems. However, this repeated independent process often leads to the same mistakes, making the selected solution still incorrect. We propose a novel prompting method named Stepwise Correction (StepCo) that helps LLMs identify and revise incorrect steps in their generated reasoning paths. It iterates verification and revision phases that employ a process-supervised verifier. The verify-then-revise process not only improves answer correctness but also reduces token consumption with fewer paths needed to generate. With StepCo, a series of LLMs demonstrate exceptional performance. Notably, using GPT-4o as the backend LLM, StepCo achieves an average accuracy of 94.1 across eight datasets, significantly outperforming the state-of-the-art Best-of-N method by +2.4, while reducing token consumption by 77.8%.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。