Composite Backdoor Attacks Against Large Language Models.

Hai Huang, Zhengyu Zhao, Michael Backes, Yun Shen, Yang Zhang

Browse the full NAACL paper archive.