SEA: Low-Resource Safety Alignment for Multimodal Large Language Models via Synthetic Embeddings

Ähnliche Einträge

• ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior
  von: Lu, Weikai, et al.
  Veröffentlicht: (2025)
• Structured Visual Narratives Undermine Safety Alignment in Multimodal Large Language Models
  von: Tan, Rui Yang, et al.
  Veröffentlicht: (2026)
• Medical MLLM is Vulnerable: Cross-Modality Jailbreak and Mismatched Attacks on Medical Multimodal Large Language Models
  von: Huang, Xijie, et al.
  Veröffentlicht: (2024)
• Beyond Text: Multimodal Jailbreaking of Vision-Language and Audio Models through Perceptually Simple Transformations
  von: Kumar, Divyanshu, et al.
  Veröffentlicht: (2025)
• Test-Time Backdoor Attacks on Multimodal Large Language Models
  von: Lu, Dong, et al.
  Veröffentlicht: (2024)