大模型自发形成类脑模块化结构，暗示智能系统可能天生需要模块分工。
Modular Cognitive Architecture Emerges in Large Language Models
- 通过46个任务分析神经元激活模式，发现大模型与人脑共享模块化组织。
- 同一认知领域任务在模型中激活重叠神经元，跨领域则激活不同区域。
- 适合对神经机制、认知科学或大模型可解释性感兴趣的读者。
人类大脑展现出显著的功能特化，不同网络分别支持语言、形式推理、心智理论和物理世界推理。这种模块化结构是智能系统必须具备的基本原则，还是仅限于生物大脑的进化偶然？本文测试大型语言模型（LLM）是否也出现类似组织——另一类通过完全不同优化过程构建的智能系统。通过对46项涵盖语言、形式推理、社会推理、物理推理四个认知领域的任务进行电路分析，我们发现大模型发展出与人脑相似的模块化架构：在人脑中由相同网络支持的任务，在模型中也招募重叠的神经元；而涉及不同网络的任务则激活不同的神经元。大脑与神经网络中模块化的收敛出现，表明其可能是智能系统的根本属性。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。