LLM Fine-Tuning: Reinforcement Learning from Human Feedback 理解 ChatGPT 的關鍵訓練階段：RLHF Feb 27．20 min read．深度學習核心觀念
