How do LLMs process syntax? Do different syntactic phenomena recruit the same model units, or do they recruit distinct model components? And do different languages rely on similar units to process the same syntactic phenomenon? Check out our new preprint (to appear at ACL 2026)!

The human brain is strikingly modular, with distinct networks for language, formal reasoning, social reasoning, and physical reasoning. Is this a fundamental principle of how intelligent systems are built, or an accident of biological evolution? In our latest preprint, we find that a similar modular organization emerges in Large Language Models, another class of intelligent system. Brains and LLMs are shaped by entirely different kinds of optimization (biological evolution vs. gradient descent). That they arrive at the same modular design anyway suggests modularity may be a fundamental property of intelligent systems. 🌐 Web: pengrui-han.github.io/LLM_Mo… 📄 Paper: pengrui-han.github.io/LLM_Mo… 💻 Code & data: github.com/Pengrui-Han/LLM_M… Using circuit analyses across 46 tasks spanning four cognitive domains, we find: 1️⃣ Tasks that draw on the same network in humans recruit overlapping units in LLMs, while tasks drawing on different networks recruit distinct units. 2️⃣ These units are causally linked to model behavior. Ablating the units critical for one domain impairs performance in that domain (−26% accuracy) but barely touches the others (−2.5%). This project has been in the works for a while :) Huge thanks to my advisors @jacobandreas @ev_fedorenko @devarda_a, and to @Nancy_Kanwisher for valuable conceptual input and feedback throughout. #MIT

The human brain is strikingly modular, with distinct networks for language, formal reasoning, social reasoning, and physical reasoning. Is this a fundamental principle of how intelligent systems are built, or an accident of biological evolution? In our latest preprint, we find that a similar modular organization emerges in Large Language Models, another class of intelligent system. Brains and LLMs are shaped by entirely different kinds of optimization (biological evolution vs. gradient descent). That they arrive at the same modular design anyway suggests modularity may be a fundamental property of intelligent systems. 🌐 Web: pengrui-han.github.io/LLM_Mo… 📄 Paper: pengrui-han.github.io/LLM_Mo… 💻 Code & data: github.com/Pengrui-Han/LLM_M… Using circuit analyses across 46 tasks spanning four cognitive domains, we find: 1️⃣ Tasks that draw on the same network in humans recruit overlapping units in LLMs, while tasks drawing on different networks recruit distinct units. 2️⃣ These units are causally linked to model behavior. Ablating the units critical for one domain impairs performance in that domain (−26% accuracy) but barely touches the others (−2.5%). This project has been in the works for a while :) Huge thanks to my advisors @jacobandreas @ev_fedorenko @devarda_a, and to @Nancy_Kanwisher for valuable conceptual input and feedback throughout. #MIT

🚨New preprint!🚨 We know that LM representations can be used to predict brain responses to language. But what *features* of these representations underlie this alignment? We use SAEs to find out!

Congratulations to Yang ICoN fellow Andrea de Varda @devarda_a, who has been named a winner of the prestigious Glushko Dissertation Prize! 👏👏 🔗 cognitivesciencesociety.org/… @cogsci_soc

Excited to share that our paper: ‘Precision fMRI reveals that the language network exhibits adult-like left-hemispheric lateralization by 4 years of age’ is finally published! mcgovern.mit.edu/2026/05/17/… nature.com/articles/s41467-0… @Amanda_M_OBrien @ev_fedorenko

Why do languages share common properties? Adults learned novel quantifiers satisfying semantic universals faster than those violating them. This suggests that learnability helps explain why certain meanings are lexicalized across cultures. @Logic_Cognition