TY  - RPRT
TI  - Knowledge Vector Weakening: Efficient Training-free Unlearning for Large Vision-Language Models
AU  - Yejin Kim
AU  - Dongjun Hwang
AU  - Sungmin Cha
AU  - Junsuk Choe
PY  - 2026
UR  - https://arxiv.org/abs/2601.21794
ID  - 2601.21794
ER  -