Earlier Tokens Contribute More: Learning Direct Preference Optimization From Temporal Decay Perspective - researchr publication related

researchr

You are not signed in
Sign in
Sign up

Ruichen Shao, Bei Li, Gangao Liu, Yang Chen, Zhouxiang, Jingang Wang, Xunliang Cai, Peng Li. Earlier Tokens Contribute More: Learning Direct Preference Optimization From Temporal Decay Perspective. In The Thirteenth International Conference on Learning Representations, ICLR 2025, Singapore, April 24-28, 2025. OpenReview.net, 2025. [doi]

The following publications are possibly variants of this publication:

Token-level Direct Preference OptimizationYongcheng Zeng, Guoqing Liu, Weiyu Ma, Ning Yang, Haifeng Zhang, Jun Wang. icml 2024: [doi]

Optimization with mode-directed preferencesHai-Feng Guo, Bharat Jayaraman, Gopal Gupta, Miao Liu. ppdp 2005: 242-251 [doi]

runs on WebDSL