On the Vulnerability of Concept Erasure in Diffusion Models

This repository contains the supplementary materials for our paper
On the Vulnerability of Concept Erasure in Diffusion Models.

Overview

This repository will eventually include all the code used for our experiments, including our adversarial prompt attack method (RECORD) and scripts for reproducing key figures and evaluation metrics from the paper. At the moment, the repository serves as a placeholder and will be updated with code once the paper is accepted.

Current Status

• Code Release: Not yet available; to be published after acceptance.
• Paper Versions: A preprint is available on arXiv (link to be added when available).

Citation

If you use or refer to this work, please cite our paper: