RL on Incorrect Synthetic Data Scales theEfficiency of LLM Math Reasoning by Eight-Fold 논문리뷰

https://arxiv.org/pdf/2406.14532