Mechanistic Interpretability for Language Models

An expert course on opening transformer internals with sparse features, circuits, attribution, interventions, and behavioral case studies covering planning, hallucination, and unfaithful reasoning.