Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selection

@article{senjie2026,
  title = {Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selection},
  author = {Jin, Senjie and Wang, Peixin and Liu, Boyang and Fan, Xiaoran and Li, Shuo and Xi, Zhiheng and Zhang, Jiazheng and Zhou, Yuhao and Gui, Tao and Zhang, Qi and Huang, Xuanjing},
  journal = {Findings of the Association for Computational Linguistics: EMNLP 2026},
  year = {2026},
  month = {Jan},
}

Jin, Senjie et al. (2026). Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selection. Findings of the Association for Computational Linguistics: EMNLP 2026.

Jin, Senjie, et al.. "Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selection." Findings of the Association for Computational Linguistics: EMNLP 2026, 2026.

Jin, Senjie, et al.. "Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selection." Findings of the Association for Computational Linguistics: EMNLP 2026 (2026).

Jin, Senjie et al., 2026. Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selection. Findings of the Association for Computational Linguistics: EMNLP 2026.

TY  - JOUR
TI  - Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selection
AU  - Jin, Senjie
AU  - Wang, Peixin
AU  - Liu, Boyang
AU  - Fan, Xiaoran
AU  - Li, Shuo
AU  - Xi, Zhiheng
AU  - Zhang, Jiazheng
AU  - Zhou, Yuhao
AU  - Gui, Tao
AU  - Zhang, Qi
AU  - Huang, Xuanjing
PY  - 2026
JO  - Findings of the Association for Computational Linguistics: EMNLP 2026
ER  -