1
Trust Region Masking for Long-Horizon LLM Reinforcement Learning
新方法Trust Region Masking，通过动态约束策略更新提升大模型长序列强化学习稳定性与收敛性。
arXiv:2512.23075v5 Announce Type: replace-cross Abstract: Policy gradient methods for Large Language Models optimize a policy $\pi_\theta$ via a surro…