Unverified50% confidenceTradeoffExact time
强化学习训练比监督学习更不稳定，将NVFP4激进量化引入RL训练会进一步放大固有不稳定性
1
Sources
50%
Confidence
Long-term
Relevance
7/14/2026
First Seen
Sources
NVFP4强化学习训练：4位量化稳定性挑战与工程实践
Related Claims
VerifiedTraditional RLHF requires first training a reward model and then optimizing with reinforcement learning algorithms such as PPO, making the process complex and unstable.77% similarUnverified近期研究表明RL（尤其是RLVR）并没有给模型注入大量新知识，而更像是激发和对齐预训练阶段已具备的能力75% similarUnverifiedRLHF和DPO等对齐技术的效果高度依赖预训练基础质量，预训练阶段未习得的能力几乎无法通过后续对齐补救73% similarUnverifiedIn the face of rapidly iterating model capabilities, establishing a continuous learning mechanism has greater strategic value than one-time deployment.72% similarUnverifiedRLHF训练中会加入KL散度惩罚项，防止模型偏离SFT阶段学到的基础行为太远70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/510048API
curl https://kongchang.com/api/v1/knowledge/claims/510048MCP
get_claim(id=510048)