// tag archive

chain-of-thought

1 article

OpenAI's o1 model uses reasoning tokens and reinforcement learning to solve competition-level geometry problems, yet struggles with the…

October 16, 2026 · 10 min read