@article{magneticpreferenceoptimizationachieving, title = {Magnetic Preference Optimization: Achieving Last-iterate Convergence for Language Model Alignment}, author = {Mingzhi Wang and Chengdong Ma and Qizhi Chen and Linjian Meng and Yang Han and Jiancong Xiao and Zhaowei Zhang and Jing Huo and Weijie J. Su and Yaodong Yang}, year = {2024}, eprint = {2410.16714}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2410.16714v3}, }