TY  - RPRT
TI  - TGDPO: Harnessing Token-Level Reward Guidance for Enhancing Direct Preference Optimization
AU  - Mingkang Zhu
AU  - Xi Chen
AU  - Zhongdao Wang
AU  - Bei Yu
AU  - Hengshuang Zhao
AU  - Jiaya Jia
PY  - 2025
UR  - https://arxiv.org/abs/2506.14574
ID  - 2506.14574
ER  -