Cascading Adversarial Bias from Injection to Distillation in Language Models.

Harsh Chaudhari, Jamie Hayes, Matthew Jagielski, Ilia Shumailov, Milad Nasr, Alina Oprea

Browse the full CCS paper archive.