Fetching the paper…

RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP Models · Around