Study Finds Modular Cognitive Networks in Large Language Models Mirror Human Brain Organization
By
Pengrui Han·Jacob Andreas·Evelina Fedorenko†·Andrea Gregor de Varda†
Summary
This research article investigates whether Large Language Models (LLMs) exhibit modular cognitive organization similar to the human brain's functional specialization. Using circuit analyses across six frontier LLMs (24B–123B parameters), the study finds that reasoning is supported by four segregated neuron populations that mirror the cognitive networks of the human brain — including networks for language, formal reasoning, reasoning about other minds (theory of mind), and reasoning about the physical world. The findings suggest that modular cognitive architecture may be a fundamental principle of intelligent systems, not merely an evolutionary accident of biological brains.
Source
Twitter / XStudy Finds Modular Cognitive Networks in Large Language Models Mirror Human Brain Organizationpengrui-han.github.ioKey quotes
· 3 pulledThe 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?
Across six frontier LLMs (24B–123B), reasoning is supported by four segregated neuron populations that mirror the cognitive networks of the human brain.
You might also wanna read
Neuroscientist Ev Fedorenko's Research on the Brain's Language Network and Its Similarities to AI Models
Neuroscientist Ev Fedorenko has spent 15 years studying the brain's language network, discovering that language processing in the human brai
quantamagazine.org·6mo ago
Enhancing Abstraction in Large Language Models Through Nature-Inspired Semantic Patterns
ieeexplore.ieee.org·1y agoThe Conceptual Challenge of Evaluating Large Language Models: When Language Fails to Describe Novel Technology
The article examines the psychological and linguistic challenges in evaluating Large Language Models (LLMs), arguing that their novel nature
LLMorphism: The biased belief that human cognition works like a large language model
This article introduces the concept of "LLMorphism" — the biased belief that human cognition works like a large language model. The author a
arXiv.org·1mo ago
Schema theory revisited: A spectrum of abstraction in AI and neuroscience
This article examines schema theory—a foundational concept in psychology, cognitive science, and neuroscience—and its relevance to both biol
Research Reveals LLMs Contain Built-In Persona Subnetworks Without External Training
This research paper reveals that large language models (LLMs) already contain specialized persona subnetworks within their parameter space,
arxiv.org·4mo ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.