BadPre: Task-agnostic Backdoor Attacks to Pre-trained NLP Foundation Models

Open in new window