大语言模型的链式推理是否是一种幻象？一种数据分布视角
Is Chain-of-Thought Reasoning of LLMs a Mirage? A Data Distribution Lens
AI总结：
本文通过数据分布视角探讨了大语言模型链式推理的有效性，揭示其在分布差异下的脆弱性，提出DataAlchemy环境验证了推理能力受训练分布影响的结论。
摘要：
链式推理（CoT）提示已被证明能有效激发大语言模型（LLMs）的结构化推理。尽管其流行，近期研究揭示其在某些推理任务中的失败，引发了对CoT推理本质的根本质疑。本文提出数据分布视角，探讨CoT推理何时何地成功或失败。我们假设CoT推理反映了一种从分布内数据学习的结构归纳偏差，使模型能够有条件生成近似训练期间观察到的推理轨迹。因此，CoT推理的有效性根本上由训练数据与测试查询之间的分布差异性质和程度决定。基于此视角，我们从任务、长度和格式三个维度剖析CoT推理。为验证假设，我们引入DataAlchemy，一个抽象且完全可控的环境，从头训练LLMs并系统地在各种分布条件下测试。通过严格控制实验，我们发现当CoT推理被推至训练分布之外时，其成为一种脆弱的幻象，强调实现真正且可推广的推理仍面临持续挑战。
英文摘要：
Chain-of-Thought (CoT) prompting has been shown to be effective in eliciting structured reasoning (i.e., CoT reasoning) from large language models (LLMs). Regardless of its popularity, recent studies expose its failures in some reasoning tasks, raising fundamental questions about the nature of CoT reasoning. In this work, we propose a data distribution lens to understand when and why CoT reasoning succeeds or fails. We hypothesize that CoT reasoning reflects a structured inductive bias learned from in-distribution data, enabling models to conditionally generate reasoning trajectories that approximate those observed during training. As such, the effectiveness of CoT reasoning is fundamentally governed by the nature and degree of distribution discrepancy between training data and test queries. Guided by this lens, we dissect CoT reasoning via three dimensions: task, length, and format. To test the hypothesis, we introduce DataAlchemy, an abstract and fully controllable environment that trains LLMs from scratch and systematically probes them under various distribution conditions. Through rigorous controlled experiments, we reveal that CoT reasoning is a brittle mirage when it is pushed beyond training distributions, emphasizing the ongoing challenge of achieving genuine and generalizable reasoning.
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。
