RLHF
概念
训练与推理
Reinforcement learning from human feedback — the technique behind ChatGPT's manners.
您正在查看一条概念解释记录：该术语的通俗含义，以及它所关联的各条目。
RLHF trains a reward model from human preference comparisons, then optimizes the language model against it. Introduced at scale by InstructGPT (2022), it is why assistants follow instructions and refuse harm — and its quirks (sycophancy, reward hacking) drive research into alternatives like constitutional AI and direct preference optimization.
它如何与整个技术栈相连
本记录的关联链接一览——点击分组可直接跳转。
来源
记录审阅截至 2026-01-15 — 查看 方法论.