This repository contains the code for a project comparing fMRI data with representations from Large Language Models.

The work provides insights into how LLMs learn during training and how they process information once trained. We did not publish this work as a paper because similar results from MIT were released shortly before we could submit.