Skip to yearly menu bar Skip to main content


Social

Mechanistic Interpretability Social

Gabriele Sarti · Nikhil Prakash

Schubert 5
[ ]
Wed 8 May 3:45 a.m. PDT — 5:15 a.m. PDT

Abstract:

Our event aims to gather researchers and practitioners interested in discussing recent advances and promising directions for deep learning interpretability research, with a specific focus on Transformers-based language models and mechanistic approaches aimed at reverse-engineering their behaviors.

Chat is not available.