Revisiting Character-level Adversarial Attacks for Language Models.

Elas Abad-Rocamora, Yongtao Wu, Fanghui Liu, Grigorios Chrysos, Volkan Cevher

Browse the full ICML paper archive.