AI RESEARCH

Erased but Not Forgotten: How Backdoors Compromise Concept Erasure

arXiv CS.AI

ArXi:2504.21072v2 Announce Type: replace-cross The expansion of text-to-image diffusion models has raised concerns about harmful outputs, from fabricated depictions of public figures to sexually explicit imagery. To mitigate such risks, prior work has proposed concept erasure methods that aim to sever unwanted concepts from the model via fine-tuning, yet it remains unclear whether these approaches truly remove all links to the harmful concept or merely conceal superficial connections.