Verified80% confidenceFactTime unknown
PPO（近端策略优化）是OpenAI在2017年提出的一种强化学习算法，成为RLHF的主流选择
4
Sources
80%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Cite This Claim
Stable URI
https://kongchang.com/claim/31969API
curl https://kongchang.com/api/v1/knowledge/claims/31969MCP
get_claim(id=31969)