A arxiv.org Smaller, Weaker, yet Better: Training LLM Reasoners via Compute-Optimal Sampling 59 points towaihee 2 years ago 5 comments
