
38.2 - Jesse Hoogland on Singular Learning Theory
AXRP - the AI X-risk Research Podcast
Exploring Multigram Circuits and Layer Coordination in Neural Models
This chapter explores the intricacies of a model's processing of n-grams and skip n-grams across its layers. It examines the impact of the learning coefficient and the coordination between heads, highlighting how predictive capabilities evolve as the model scales.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.