Keeping AI Under Control Through Mechanistic Interpretability (MIT Talk)youtube.com5 points·Lordarminius··0 commentsOpen articleSaveView on HN