Bilal Chughtai

This review discusses the current frontier of mechanistic interpretability, which aims to understand the computational mechanisms underlying neural networks. While the field has made progress, many open problems remain, including the need for improved methods, better applications to specific goals, and engagement with socio-technical challenges.

Existing circuits in the mechanistic interpretability literature may not be as faithful as reported. Current circuit faithfulness scores reflect both the methodological choices of researchers and the actual components of the circuit.