Fetching the paper…

Optimizing Chain-of-Thought Reasoners via Gradient Variance Minimization in Rejection Sampling and RL · Around