Defending Pre-trained Language Models as Few-shot Learners against Backdoor Attacks

Devs

Defending Pre-trained Language Models as Few-shot Learners against Backdoor Attacks | Read Paper on Bytez