MIT Researchers Read and Manipulate LLMs' Internal Physics Intuition
MIT's Markus Buehler team published in AI for Science showing that interpretability methods can read and manipulate LLMs' internal representations of materials science physics, using open-weight Gemma models (4B/12B/31B).
2026-11-19 ~ 2026-11-19 · 2 related posts
- MIT reads and steers LLMs' internal physics intuition via hidden-state shifts in Gemma — ProfBuehlerMIT · 2026-11-19
- MIT's Buehler reads and steers materials-science physics representations inside open-weight Gemma 4 LLMs — ProfBuehlerMIT · 2026-11-19