BackdoorHunter: Poisoned Training Data Removal via Contrastive Responses Under Pruned Models.

Rixi Liang, Shuai Zhou, Mingxu Zhu, Chi Liu, Minfeng Qi

Browse the full KSEM paper archive.