InstructTA: Instruction-Tuned Targeted Attack for Large Vision-Language Models

MLLMRe-assign community