Similar to エンドツーエンドの強化学習

Reinforcement learning from human feedback

Proximal policy optimization