@article{layerlevelselfexposureandpatch, title = {Layer-Level Self-Exposure and Patch: Affirmative Token Mitigation for Jailbreak Attack Defense}, author = {Yang Ouyang and Hengrui Gu and Shuhang Lin and Wenyue Hua and Jie Peng and Bhavya Kailkhura and Meijun Gao and Tianlong Chen and Kaixiong Zhou}, year = {2025}, eprint = {2501.02629}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2501.02629v2}, }