Reinforcement Learning from Human Feedback (RLHF)

Improving chatbot tone through user ratings.

Synonyms:

Human-in-the-loop reinforcement learning, Preference-based RL

Defnition:
RLHF trains AI models using human feedback to improve alignment and output quality.

Variations:
LLM alignment training using human feedback