数学拿A：渐进式纠错提示法
Get an A in Math: Progressive Rectification Prompting
AI总结：
针对思维链提示易因推理路径错误导致数学应用题答案不准的问题，提出渐进式纠错提示法PRP，通过迭代验证纠错提升准确率，在8个数据集上从77.3%升至90.5%，性能优于现有思维链方法。
摘要：
思维链（Chain-of-Thought, CoT）提示方法已使大语言模型（LLM）能够生成推理路径并解决数学应用题（MWP）。然而，这类方法对推理路径中的错误十分敏感，任何错误都可能导致最终答案不正确。我们提出了一种名为Progressive Rectification Prompting（PRP，渐进式纠错提示）的新方法，将8个MWP数据集上的平均准确率从77.3提升至90.5。在获得CoT生成的初始答案后，PRP会迭代执行“验证-纠错”流程，逐步识别错误答案并修正推理路径。针对可能性最高的正确答案，LLM会预测问题中被掩码的数值；若预测值与掩码值不匹配，则该答案大概率是错误的。随后，系统会提示LLM结合一组错误答案的提示重新生成推理路径，避免重复之前的错误。与各类CoT方法相比，PRP取得了最优性能。我们的实现代码已公开于https://wzy6642.github.io/prp.github.io/。
英文摘要：
Chain-of-Thought (CoT) prompting methods have enabled large language models (LLMs) to generate reasoning paths and solve math word problems (MWPs). However, they are sensitive to mistakes in the paths, as any mistake can result in an incorrect answer. We propose a novel method named Progressive Rectification Prompting (PRP) to improve average accuracy on eight MWP datasets from 77.3 to 90.5. Given an initial answer from CoT, PRP iterates a verify-then-rectify process to progressively identify incorrect answers and rectify the reasoning paths. With the most likely correct answer, the LLM predicts a masked numerical value in the question; if the prediction does not match the masked value, the answer is likely incorrect. Then the LLM is prompted to re-generate the reasoning path hinted with a set of incorrect answers to prevent itself from repeating previous mistakes. PRP achieves the best performance compared against the CoT methods. Our implementation is made publicly available at https://wzy6642.github.io/prp.github.io/.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
