Mechanistic Interpretability Benchmark

AI & ML interests

Principled evaluation of mechanistic interpretability methods.

Recent Activity

hadasor  authored a paper about 1 month ago

Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types

hadasor  authored a paper about 1 month ago

Agents of Chaos

hadasor  authored a paper about 1 month ago

Inside-Out: Hidden Factual Knowledge in LLMs

View all activity

Team members 20