Our paper “Imperio: Language-Guided Backdoor Attacks for Arbitrary Model Control” has been accepted to IJCAI 2024. [PDF] [Code]

Published: April 15, 2024