baobab.bio
We build biological foundation models that are interpretable and performant.
Project 1: A DNA language model with many interpretable neurons – preprint.
Research note: A very simple, very useful interpretability technique for protein language models.