Distinguishing Non-natural from Natural Adversarial Samples for More Robust Pre-trained Language Model.

Jiayi Wang, Rongzhou Bao, Zhuosheng Zhang, Hai Zhao

Browse the full ACL paper archive.