JiazhaoLi / AttDef

Dataset and Code for Findings of ACL 2023 paper: Defending against Insertion-based Textual Backdoor Attacks via Attribution
4 stars 0 forks source link

运行脚本问题咨询 #1

Open wellzhi opened 7 months ago

wellzhi commented 7 months ago
image

作者你好,脚本此处缺失,请问怎么处理,谢谢

JiazhaoLi commented 7 months ago

你好,这里我们设置 将 poisoned model 存在哪里,你可以根据你的实际情况 自己设置 target path.