1 paper
Chuancheng Shi, Shangze Li, Wenhua Wu +5
Despite their remarkable capabilities, large foundation models (LFMs) remain susceptible to adversarial manipulation. Current safety intervention methods primarily rely on the \tex…