That is, the paper posits the mechanism isn't branch prediction but multi-scale learning. Since the correct parts of the chain of thought remain more frequent than the incorrect parts they get…

Posters: jd_pressman. jd_pressman

2 likes recorded by the source corpus.