Author name: Wei Chen, Yubing Wu, Junmei Yang, Delu Zeng, Qibin Zhao, John Paisley, Min Chen, Zhou Wang

arXiv:2604.18239v1 Announce Type: new
Abstract: Preference optimization is widely used to align large language models (LLMs) with human preferences. However, many margin-based objectives suppress the chosen response along with the rejected one, a phen…