Defending Language Models Against Image-Based Prompt Attacks via User-Provided Specifications.

Reshabh K. Sharma, Vinayak Gupta, Dan Grossman

Browse the full SP paper archive.