大型语言模型中出现的模块化认知架构
Modular Cognitive Architecture Emerges in Large Language Models
浏览论文内容
中文总结 AI 辅助
该研究探究大型语言模型是否会出现类似人类大脑的模块化认知架构，经对4个认知领域46项任务的回路分析，发现其会形成相似模块化架构，表明模块化可能是智能系统的基本属性。
中文摘要 AI 辅助
人类大脑展现出显著的功能特化程度，不同的网络分别支持语言、形式推理、对他人心智的推理以及对物理世界的推理。这种模块化组织是智能系统构建必须遵循的基本原理，还是仅属于生物大脑的进化偶然？本研究测试大型语言模型（另一类通过完全不同优化过程生成的智能系统）是否会出现类似组织。通过对涵盖语言、形式推理、社会推理、物理推理四个认知领域的N=46项任务进行回路分析，研究发现大型语言模型会形成与人类大脑相似的模块化架构：依赖人类同一网络的任务会招募大型语言模型中重叠的神经元，而依赖人类不同网络的任务则会招募不同的神经元。大脑与神经网络中模块化的趋同出现表明，模块化可能是智能系统的一种基本属性。
英文摘要
The human brain exhibits a striking degree of functional specialization, with distinct networks supporting language, formal reasoning, reasoning about other minds, and reasoning about the physical world. Is this modular organization a fundamental principle of how intelligent systems must be built, or an evolutionary accident specific to biological brains? Here, we test whether a similar organization emerges in Large Language Models--another class of intelligent systems created through a very different optimization process. Using circuit analyses across N=46 tasks spanning four cognitive domains (language, formal reasoning, social reasoning, physical reasoning), we find that LLMs develop a modular architecture that mirrors the human brain: tasks drawing on the same network in humans recruit overlapping neurons in LLMs, whereas tasks drawing on different networks recruit distinct neurons. The convergent emergence of modularity in brains and neural networks suggests that it may be a fundamental property of intelligent systems.
