This page may contain stale information. Last updated: 2026-09-19
Definition
RLHF aligns language models to human preferences via reward modeling — foundational to ChatGPT-style assistants. diogo-almeida co-invented RLHF at OpenAI before founding typesafe-ai.