@article{activationsteeringmeetspreferenceoptimiz, title = {Activation Steering Meets Preference Optimization: Defense Against Jailbreaks in Vision Language Models}, author = {Sihao Wu and Gaojie Jin and Wei Huang and Jianhong Wang and Xiaowei Huang}, year = {2025}, eprint = {2509.00373}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2509.00373}, }