Researchers just got very specific about what a language model’s ‘chain of thought’ is actually doing.

Aller à la source