从霍普菲尔德视角理解思维链中的推理
Understanding Reasoning in Chain-of-Thought from the Hopfieldian View
AI总结：
本文从霍普菲尔德认知观出发，将思维链推理建模为表征空间中的移动，并提出思维表征（RoT）框架，以提升推理的鲁棒性与可解释性。
摘要：
大型语言模型已在各种任务中展现出卓越能力，其中思维链（CoT）提示已成为增强推理能力的关键技术。然而，现有研究主要侧重于提升性能，缺乏一个全面的框架来解释和理解CoT成功背后的根本因素。为弥合这一差距，我们引入了一种基于认知神经科学中霍普菲尔德认知观的新视角。我们在CoT推理与刺激、行为、神经群体和表征空间等关键认知要素之间建立了联系。从我们的视角出发，可以将推理过程理解为在这些表征空间之间的移动。基于这一见解，我们开发了一种在CoT响应中定位推理错误的方法。此外，我们提出了思维表征（Representation-of-Thought，RoT）框架，该框架利用低维表征空间的鲁棒性来增强CoT中推理过程的鲁棒性。实验结果表明，RoT提高了CoT推理的鲁棒性和可解释性，同时提供了对推理过程的细粒度控制。
英文摘要：
Large Language Models have demonstrated remarkable abilities across various tasks, with Chain-of-Thought (CoT) prompting emerging as a key technique to enhance reasoning capabilities. However, existing research primarily focuses on improving performance, lacking a comprehensive framework to explain and understand the fundamental factors behind CoT's success. To bridge this gap, we introduce a novel perspective grounded in the Hopfieldian view of cognition in cognitive neuroscience. We establish a connection between CoT reasoning and key cognitive elements such as stimuli, actions, neural populations, and representation spaces. From our view, we can understand the reasoning process as the movement between these representation spaces. Building on this insight, we develop a method for localizing reasoning errors in the response of CoTs. Moreover, we propose the Representation-of-Thought (RoT) framework, which leverages the robustness of low-dimensional representation spaces to enhance the robustness of the reasoning process in CoTs. Experimental results demonstrate that RoT improves the robustness and interpretability of CoT reasoning while offering fine-grained control over the reasoning process.
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。
