#alignment

RLHF: Reinforcement Learning from Human Feedback

Teaching language models to prefer responses that people rank higher