Unverified50% confidenceFactExact time
GRPO由DeepSeek团队系统化提出并在训练DeepSeek-R1系列模型时得到大规模验证
1
Sources
50%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
Related Claims
VerifiedDeepSeek-R1通过在代码和数学两个可验证域上大规模运行GRPO算法取得成功80% similarUnverifiedDeepSeek extensively uses the GRPO algorithm in place of traditional RLHF during the training phase to reduce training costs.80% similarVerifiedGRPO (Group Relative Policy Optimization), proposed by DeepSeek, requires no separate training of a value network and has become the mainstream algorithm choice for current agentic RL training.80% similarUnverifiedDeepSeek explored the GRPO algorithm as an attempt to improve upon the original RLHF framework.79% similarUnverifiedDeepSeek团队报告，仅用GRPO在数学和代码数据上进行强化学习而不经过监督微调，模型便自发出现反思、回溯验证等链式推理行为（aha moment）77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/461042API
curl https://kongchang.com/api/v1/knowledge/claims/461042MCP
get_claim(id=461042)