Foundations
Transformers, attention and representations
TRACK B · RESEARCH
Inside AI
Study transformers, internal representations, circuits, sparse autoencoders and mechanistic interpretability.
Learning themes
Mechanistic analysis
Circuits, features and sparse autoencoders
Research practice
Interventions, ablations and reproducibility
Lessons will be released after their full materials, experiments and review criteria are ready. Planned topics are not presented as published courses.
First lesson · Theme B.1
B.1.1 · What Is Mechanistic Interpretability?
Open Russian preview ↗The English lesson translation is in preparation. The current Russian draft is available as a clearly labelled preview.
EN TRANSLATION IN PREPARATION · NOT PUBLISHED