PPO

RLHF Explained: How Human Feedback Steers Reinforcement Learning

A clear, practical guide to RLHF—how human preferences train models, the pipeline, pitfalls, and modern variants like DPO and RLAIF.