大型语言模型能否检测长思维链推理中的错误？
Can Large Language Models Detect Errors in Long Chain-of-Thought Reasoning?
AI总结：
为评估大型语言模型对长思维链推理的批判能力，本文提出DeltaBench基准，细粒度分析不同o1-like模型生成的长CoT，并广泛评估现有PRMs和批判模型的错误检测边界与局限。
摘要：
近期，o1-like模型引起了广泛关注，这些模型通过生成较长的Chain-of-Thought (CoT)推理步骤来提升现有大型语言模型（LLMs）的推理能力。为了理解这些长CoT的质量并衡量现有LLMs对这些长CoT的批判能力，我们提出了DeltaBench。该基准包含来自不同o1-like模型（如QwQ、DeepSeek-R1）针对不同推理任务（如数学、代码、通用推理）生成的长CoT，用于衡量检测长CoT推理中错误的能力。基于DeltaBench，我们首先对生成的长CoT进行细粒度分析，以探究不同o1-like模型的有效性与效率。随后，我们对现有的过程奖励模型（PRMs）和批判模型进行广泛评估，以检测每个标注过程中的错误，旨在调查现有PRMs和批判模型的边界与局限性。最后，我们希望DeltaBench能指导开发者更好地理解其模型的长CoT推理能力。
英文摘要：
Recently, o1-like models have drawn significant attention, where these models produce the long Chain-of-Thought (CoT) reasoning steps to improve the reasoning abilities of existing Large Language Models (LLMs). In this paper, to understand the qualities of these long CoTs and measure the critique abilities of existing LLMs on these long CoTs, we introduce the DeltaBench, including the generated long CoTs from different o1-like models (e.g., QwQ, DeepSeek-R1) for different reasoning tasks (e.g., Math, Code, General Reasoning), to measure the ability to detect errors in long CoT reasoning. Based on DeltaBench, we first perform fine-grained analysis of the generated long CoTs to discover the effectiveness and efficiency of different o1-like models. Then, we conduct extensive evaluations of existing process reward models (PRMs) and critic models to detect the errors of each annotated process, which aims to investigate the boundaries and limitations of existing PRMs and critic models. Finally, we hope that DeltaBench could guide developers to better understand the long CoT reasoning abilities of their models.
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。
