METR
· Tech Media
Fine-tuning experiments on CoT controllability
p:has(> img) { margin-bottom: 0; } .content img { margin: 0.75em 0; } Kei Nishimura-Gasparian is an Astra fellow and was the primary contributor to this work. Neev Parikh provided mentorship and feedback. Summary: We find that a small amount of fine-tuning on instruction following in the CoT generalizes to meaningful i