PADD: Prefix-based Attention Divergence Detector for LLM Jailbreaks.

@inproceedings{bao-2026-padd, author = {Ziqun Bao and Jiaqiang Niu and Yuchen Shao and Chengcheng Wan}, editor = {Hakim Hacid and Yoelle Maarek and Francesco Bonchi and Ido Guy and Emine Yilmaz}, title = {{PADD: Prefix-based Attention Divergence Detector for LLM Jailbreaks}}, booktitle = {{Proceedings of the ACM Web Conference 2026, WWW 2026, Dubai, United Arab Emirates, originally scheduled for April 13-17, 2026, rescheduled for June 29 - July 3, 2026}}, pages = {2683--2694}, publisher = {ACM}, year = {2026}, url = {https://doi.org/10.1145/3774904.3792204}, doi = {https://doi.org/10.1145/3774904.3792204} }