Backdoor Attacks for In-Context Learning with Language Models. (arXiv:2307.14692v1 [cs.CR])

Because state-of-the-art language models are expensive to train, most
practitioners must make use of one of the few publicly available language
models or language model APIs. This consolidation of trust increases the
potency of backdoor attacks, where an adversary tampers with a machine learning
model in order to make it perform some malicious behavior on inputs that
contain a predefined backdoor trigger. We show that the in-context learning
ability of large language models significantly complicates the question of
developing backdoor …

adversary apis art attacks backdoor backdoor attacks consolidation context language language models machine machine learning order state train trust