Exploiting Prompt-induced Confidence for Black-Box Attacks on LLMs.

Meina Chen, Yihong Tang, Kehai Chen

Browse the full EMNLP paper archive.