1
VRPO: Rethinking Value Modeling for Robust RL under Noisy Supervision in LLM Post-Training
应对LLM后训练中噪声监督，VRPO提出重新思考价值建模实现鲁棒强化学习，值得关注。
arXiv:2508.03058v2 Announce Type: replace-cross Abstract: Reinforcement Learning (RL) in real-world environments often suffers from ambiguous or incom…