My biggest R1 update is I suspect <thinking> fixes depth deficiencies by giving more passes over the residual stream. Easy to forget that even @teortaxestex was bummed about small model smell…

Posters: georgejrjrjr. georgejrjrjr

2 likes recorded by the source corpus.

Topics: AI & LLMs

View Post: https://x.com/georgejrjrjr/status/1886965092131975512