@misc{indiciaea90ad3504124,
  title = {Prefill-level Jailbreak: A Black-Box Risk Analysis of Large Language Models},
  author = {Yakai Li and Jiekang Hu and Weiduan Sang and Luping Ma and Dongsheng Nie and Weijuan Zhang and Aimin Yu and Yi Su and Qingjia Huang and Qihang Zhou},
  year = {2025},
  doi = {10.1109/dsn69566.2026.00054},
  url = {https://arxiv.org/abs/2504.21038},
  note = {Source identifier: 2504.21038}
}