@misc{indiciaeb2b99d7c5941,
  title = {Not All Tokens See Equally: Perception-Grounded Policy Optimization for Large Vision-Language Models},
  author = {Zekai Ye and Qiming Li and Xiaocheng Feng and Ruihan Chen and Ziming Li and Haoyu Ren and Kun Chen and Dandan Tu and Bing Qin},
  year = {2026},
  url = {https://arxiv.org/abs/2604.01840},
  note = {Source identifier: 2604.01840}
}