Training Language Models to Self-Correct via Reinforcement Learning

Kumar, Aviral, Zhuang, Vincent, Agarwal, Rishabh · arXiv · 2024