GenErase：扩散模型中可泛化且语义感知的概念擦除
GenErase: Generalizable and Semantically-Aware Concept Erasure in Diffusion Models
在交叉注意力 value 空间做显式 erase-and-replace 并配合 per-token preserve projector 与硬几何门控，是一种可直接复用的免训练概念抑制手段，可用于个性化生成中去除身份/风格/物体，同时保护无关语义不被扭曲。
原文依据 GenErase enforces semantic orthogonality in the cross-attention value space via an explicit erase-and-replace operation
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论文本到图像（T2I）扩散模型支撑着现代创意工具，但其开放式生成特性引发了安全、伦理与版权方面的担忧。为移除每一个不安全或受版权保护的概念而重新训练或微调并不现实，这促使人们采用免训练干预手段，在抑制特定语义的同时保持整体视觉质量。现有护栏方法面临一个核心权衡：它们要么过于刚性，无法泛化到改写或语境偏移的提示，要么过于粗糙，扭曲了无关内容与保真度。我们提出 GenErase（GENeralizable ERAsure with SEmantic Awareness），一个免训练、基于几何的框架，用于在扩散模型中实现稳健的概念移除。GenErase 通过显式的擦除—替换操作，在交叉注意力 value 空间中强制语义正交性，并由逐 token 的保留投影器与硬几何门控引导。该设计实现了精确擦除、对关键语义的显式保护，以及在跨层、改写与多概念情形下的稳定性。在身份、物体与风格擦除上的大量实验，连同新的 GenBench-40 基准，表明 GenErase 达到了最优的擦除保真度与更优的改写级泛化能力，使其成为面向安全、实时扩散部署的实用且有原则的护栏。
英文原文摘要 1359 字符
Text-to-Image (T2I) diffusion models power modern creative tools, but their open-ended generative nature raises safety, ethical, and copyright concerns. Retraining or fine-tuning to remove every unsafe or copyrighted concept is impractical, motivating training-free interventions that suppress specific semantics while preserving general visual quality. Existing guard-railing methods face a core trade-off: they are either rigid, failing to generalize to paraphrased or context-shifted prompts, or coarse, distorting unrelated content and fidelity. We present GenErase (GENeralizable ERAsure with SEmantic Awareness), a training-free, geometry-grounded framework for robust concept removal in diffusion models. GenErase enforces semantic orthogonality in the cross-attention value space via an explicit erase-and-replace operation, guided by a per-token preserve projector and a hard geometric gate. This design enables precise erasure, explicit protection of critical semantics, and stability across layers, paraphrases, and multi-concept cases. Extensive experiments on identity, object, and style erasure, together with a new GenBench-40 benchmark, show that GenErase achieves state-of-the-art erasure fidelity and superior paraphrase-level generalization, establishing it as a practical and principled guard-rail for safe, real-time diffusion deployment.