Knowledge Vector Weakening: Efficient Training-free Unlearning for Large Vision-Language Models

MUVLMRe-assign community