Peng, Shangpin

2 publications

ICLR 2026 Uni-DPO: A Unified Paradigm for Dynamic Preference Optimization of LLMs Shangpin Peng, Weinong Wang, Zhuotao Tian, Senqiao Yang, Xing W, Haotian Xu, Chengquan Zhang, Takashi Isobe, Baotian Hu, Min Zhang

ICCV 2025 Mitigating Object Hallucinations via Sentence-Level Early Intervention Shangpin Peng, Senqiao Yang, Li Jiang, Zhuotao Tian