Beyond Uniform Token-Level Trust Region in LLM Reinforcement Learning

Renjie Mao, Xiangxin Zhou, Lvfang Tao, Yixin Ding, Yu Shi, Yongguang Lin, Yuheng Wu, Honglin Zhu, Qian Qiu, Wenxi ZhuarXiv 2026 · DOI

BibTeX
@misc{mao2026beyond,
  author = {Renjie Mao and Xiangxin Zhou and Lvfang Tao and Yixin Ding and Yu Shi and Yongguang Lin and Yuheng Wu and Honglin Zhu and Qian Qiu and Wenxi Zhu},
  title = {{Beyond Uniform Token-Level Trust Region in LLM Reinforcement Learning}},
  year = {2026},
  doi = {10.48550/arxiv.2606.10968},
}

Loading related content...

Figure