[KongchangAI]
Unverified50% confidenceFactExact time

GRPO是DeepSeek团队提出的强化学习算法,它取消了PPO中的Critic Model,通过对同一问题生成一组回答计算组内相对奖励作为基线

1
Sources
50%
Confidence
Long-term
Relevance
8/6/2026
First Seen

Sources

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/697161
API
curl https://kongchang.com/api/v1/knowledge/claims/697161
MCP
get_claim(id=697161)