Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models

Add code
Jan 20, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: