FG-PRM：语言模型数学推理中的细粒度幻觉检测与缓解
FG-PRM: Fine-grained Hallucination Detection and Mitigation in Language Model Mathematical Reasoning
AI总结：
本文提出FG-PRM模型，通过六类幻觉分类体系和自动数据生成方法，在步骤级细粒度检测与缓解LLM数学推理中的幻觉，显著提升了模型在GSM8K和MATH基准上的可靠性与可解释性。
摘要：
大型语言模型（LLM）在需要复杂多步推理的任务（如数学问题求解）中存在的幻觉带来了重大挑战。现有方法主要检测幻觉的存在，但缺乏对其类型和表现的细致理解。本文首先引入一个全面的分类体系，将数学推理任务中的常见幻觉分为六类。随后提出FG-PRM（Fine-Grained Process Reward Model，细粒度过程奖励模型），这是一种增强型模型，旨在以细粒度的步骤级方式检测和缓解幻觉。为解决人工标注训练数据的局限，我们提出一种利用LLM自动生成细粒度幻觉数据的方法。FG-PRM在两项关键任务上表现卓越：1）细粒度幻觉检测：对每个推理步骤的幻觉类型进行分类；2）验证：对多个LLM生成的输出进行排序以选出最准确的解答。实验表明，FG-PRM在细粒度幻觉检测上表现优异，并大幅提升了LLM在GSM8K和MATH基准上的性能。这些结果突显了细粒度监督在增强LLM推理过程的可靠性和可解释性方面的优势。
英文摘要：
Hallucinations in large language models (LLMs) pose significant challenges in tasks requiring complex multi-step reasoning, such as mathematical problem-solving. Existing approaches primarily detect the presence of hallucinations but lack a nuanced understanding of their types and manifestations. In this paper, we first introduce a comprehensive taxonomy that categorizes the common hallucinations in mathematical reasoning tasks into six types. We then propose FG-PRM (Fine-Grained Process Reward Model), an augmented model designed to detect and mitigate hallucinations in a fine-grained, step-level manner. To address the limitations of manually labeling training data, we propose an automated method for generating fine-grained hallucination data using LLMs. Our FG-PRM demonstrates superior performance across two key tasks: 1) Fine-grained hallucination detection: classifying hallucination types for each reasoning step; and 2) Verification: ranking multiple LLM-generated outputs to select the most accurate solution. Our experiments show that FG-PRM excels in fine-grained hallucination detection and substantially boosts the performance of LLMs on GSM8K and MATH benchmarks. These results highlight the benefits of fine-grained supervision in enhancing the reliability and interpretability of LLM reasoning processes.
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。
