Login / Signup

Efficient LLM-Jailbreaking by Introducing Visual Modality.

Zhenxing NiuYuyao SunHaodong RenHaoxuan JiQuan WangXiaoke MaGang HuaRong Jin
Published in: CoRR (2024)
Keyphrases
  • multi modal
  • visual features
  • computationally expensive
  • databases
  • real world
  • learning algorithm
  • knowledge base
  • high level
  • object recognition
  • input image
  • lightweight
  • multimedia data