Chain-of-Thought (CoT) prompting has boosted the multi-step reasoning capabilities of Large Language Models (LLMs) by generating a series of rationales before the final answer. We analyze the reasoning paths generated by CoT and find two issues in multi-step reasoning: (i) Generating rationales irrelevant to the question, (ii) Unable to compose subquestions or queries for generating/retrieving all the relevant information. To address them, we propose a graph-guided CoT prompting method, which guides the LLMs to reach the correct answer with graph representation/verification steps. Specifically, we first leverage LLMs to construct a "question/rationale graph" by using knowledge extraction prompting given the initial question and the rationales generated in the previous steps. Then, the graph verification step diagnoses the current rationale triplet by comparing it with the existing question/rationale graph to filter out irrelevant rationales and generate follow-up questions to obtain relevant information. Additionally, we generate CoT paths that exclude the extracted graph information to represent the context information missed from the graph extraction. Our graph-guided reasoning method shows superior performance compared to previous CoT prompting and the variants on multi-hop question answering benchmark datasets.

翻译：思维链（Chain-of-Thought, CoT）提示通过生成最终答案前的一系列推理过程，提升了大型语言模型（LLMs）的多步推理能力。我们分析了CoT生成的推理路径，发现多步推理中存在两个问题：（i）生成与问题无关的推理依据；（ii）无法组合子问题或查询以生成/检索所有相关信息。为解决这些问题，我们提出了一种图引导的CoT提示方法，通过图表示/验证步骤引导LLMs得出正确答案。具体而言，我们首先利用LLMs，通过知识抽取提示构建"问题/推理依据图"，其中包含初始问题及前序步骤生成的推理依据。随后，图验证步骤将当前推理三元组与已有问题/推理依据图进行比对，滤除无关推理依据并生成后续问题以获取相关信息。此外，我们生成排除已抽取图信息的CoT路径，以表征图抽取中遗漏的上下文信息。实验表明，在多跳问答基准数据集上，我们的图引导推理方法相较于先前CoT提示及其变体方法展现出更优性能。