Mechanistic Interpretability Benchmark

AI & ML interests

Principled evaluation of mechanistic interpretability methods.

Recent Activity

amueller  updated a Space 6 days ago

mib-bench/leaderboard

hij  authored a paper about 2 months ago

AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders

hij  authored a paper about 2 months ago

LLMs Encode Harmfulness and Refusal Separately

View all activity

Team members 19