Erased, But Not Forgotten: Erased Rectified Flow Transformers Still Remain Unsafe Under Concept Attack
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917388744654848 |
|---|---|
| author | Jiang, Nanxiang Fan, Zhaoxin Kang, Enhan Gao, Daiheng Zhou, Yun Chang, Yanxia Zhu, Zheng Jin, Yeying Wu, Wenjun |
| author_facet | Jiang, Nanxiang Fan, Zhaoxin Kang, Enhan Gao, Daiheng Zhou, Yun Chang, Yanxia Zhu, Zheng Jin, Yeying Wu, Wenjun |
| contents | Recent advances in text-to-image (T2I) diffusion models have enabled impressive generative capabilities, but they also raise significant safety concerns due to the potential to produce harmful or undesirable content. While concept erasure has been explored as a mitigation strategy, most existing approaches and corresponding attack evaluations are tailored to Stable Diffusion (SD) and exhibit limited effectiveness when transferred to next-generation rectified flow transformers such as Flux. In this work, we present ReFlux, the first concept attack method specifically designed to assess the robustness of concept erasure in the latest rectified flow-based T2I framework. Our approach is motivated by the observation that existing concept erasure techniques, when applied to Flux, fundamentally rely on a phenomenon known as attention localization. Building on this insight, we propose a simple yet effective attack strategy that specifically targets this property. At its core, a reverse-attention optimization strategy is introduced to effectively reactivate suppressed signals while stabilizing attention. This is further reinforced by a velocity-guided dynamic that enhances the robustness of concept reactivation by steering the flow matching process, and a consistency-preserving objective that maintains the global layout and preserves unrelated content. Extensive experiments consistently demonstrate the effectiveness and efficiency of the proposed attack method, establishing a reliable benchmark for evaluating the robustness of concept erasure strategies in rectified flow transformers. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_00635 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Erased, But Not Forgotten: Erased Rectified Flow Transformers Still Remain Unsafe Under Concept Attack Jiang, Nanxiang Fan, Zhaoxin Kang, Enhan Gao, Daiheng Zhou, Yun Chang, Yanxia Zhu, Zheng Jin, Yeying Wu, Wenjun Computer Vision and Pattern Recognition Recent advances in text-to-image (T2I) diffusion models have enabled impressive generative capabilities, but they also raise significant safety concerns due to the potential to produce harmful or undesirable content. While concept erasure has been explored as a mitigation strategy, most existing approaches and corresponding attack evaluations are tailored to Stable Diffusion (SD) and exhibit limited effectiveness when transferred to next-generation rectified flow transformers such as Flux. In this work, we present ReFlux, the first concept attack method specifically designed to assess the robustness of concept erasure in the latest rectified flow-based T2I framework. Our approach is motivated by the observation that existing concept erasure techniques, when applied to Flux, fundamentally rely on a phenomenon known as attention localization. Building on this insight, we propose a simple yet effective attack strategy that specifically targets this property. At its core, a reverse-attention optimization strategy is introduced to effectively reactivate suppressed signals while stabilizing attention. This is further reinforced by a velocity-guided dynamic that enhances the robustness of concept reactivation by steering the flow matching process, and a consistency-preserving objective that maintains the global layout and preserves unrelated content. Extensive experiments consistently demonstrate the effectiveness and efficiency of the proposed attack method, establishing a reliable benchmark for evaluating the robustness of concept erasure strategies in rectified flow transformers. |
| title | Erased, But Not Forgotten: Erased Rectified Flow Transformers Still Remain Unsafe Under Concept Attack |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.00635 |