Naomi Saphra

A natural language processing researcher who studies how language models change during training, connecting optimization behavior to the generalizations models end up making. Her work draws on linguistics, deep learning theory, and interpretability, and extends to applying interpretability methods to models used for scientific discovery.