SAUCE: Selective Concept Unlearning in Vision-Language Models with Sparse Autoencoders.

Jiahui Geng, Qing Li

Browse the full ICCV paper archive.