@article{safebehaviorsimulatinghumanlikemultistag,
  title = {SafeBehavior: Simulating Human-Like Multistage Reasoning to Mitigate Jailbreak Attacks in Large Language Models},
  author = {Qinjian Zhao and Jiaqi Wang and Zhiqiang Gao and Zhihao Dou and Belal Abuhaija and Kaizhu Huang},
  year = {2025},
  eprint = {2509.26345},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2509.26345},
}