Diffusion Theory as a Scalpel: Detecting and Purifying Poisonous Dimensions in Pre-trained Language Models Caused by Backdoor or Bias

Diffusion Theory as a Scalpel: Detecting and Purifying Poisonous Dimensions in Pre-trained Language Models Caused by Backdoor or Bias Zhiyuan Zhang author Deli Chen author Hao Zhou author Fandong Meng author Jie Zhou author Xu Sun author 2023-07 text Findings of the Association for Computational Linguistics: ACL 2023 Anna Rogers editor Jordan Boyd-Graber editor Naoaki Okazaki editor Association for Computational Linguistics Toronto, Canada conference publication zhang-etal-2023-diffusion 10.18653/v1/2023.findings-acl.157 https://aclanthology.org/2023.findings-acl.157/ 2023-07 2495 2517