DMSD: Dual-Modal Safe Decoding Against White-Box and Black-Box Jailbreaks in Large Vision-Language Models
IEEE Access · 2026 · 14 · 169994-170006
DOI: 10.1109/access.2026.3745155 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Hui Ye und 4 weiteren Autor:innen. Die Arbeit nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Social Science Fund of China.
Abstract (Originalsprache)
Kein Abstract verfügbar.
Förderung
- National Social Science Fund of China
Zitieren
APA
Ye, H., Pan, T., Wang, Z., Li, H., & Zheng, L. (2026). DMSD: Dual-Modal Safe Decoding Against White-Box and Black-Box Jailbreaks in Large Vision-Language Models. IEEE Access, 14, 169994-170006. https://doi.org/10.1109/access.2026.3745155
Vancouver
Ye H, Pan T, Wang Z, Li H, Zheng L. DMSD: Dual-Modal Safe Decoding Against White-Box and Black-Box Jailbreaks in Large Vision-Language Models. IEEE Access. 2026;14:169994-170006. doi:10.1109/access.2026.3745155
BibTeX
@article{Ye2026,
author = {Ye, Hui and Pan, Tiejun and Wang, Zhang and Li, Hongbo and Zheng, Leina},
title = {DMSD: Dual-Modal Safe Decoding Against White-Box and Black-Box Jailbreaks in Large Vision-Language Models},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {169994--170006},
doi = {10.1109/access.2026.3745155}
}
EUVIMED