BadPrompt: Backdoor Attacks on Continuous Prompts.

Xiangrui Cai Haidong Xu Sihan Xu Ying Zhang Xiaojie Yuan

Published in: CoRR (2022)

Keyphrases

security threats
neural network
mental models
countermeasures
malicious attacks
data mining
continuous domains
computer security
digital watermarking
malicious users
chosen plaintext