用动态选锚技术精准擦除图像中的概念，避免复现和误伤。
Beyond Fixed Anchors: Precisely Erasing Concepts with Sibling Exclusive Counterparts
- 通过因果追踪发现最佳锚点，动态选择不干扰其他概念的擦除靶点。
- 单概念锚点挖掘仅需平均4秒，擦除后相关概念保留率显著提升。
- 适合作为通用插件集成到各类文本生成模型中，提升可控性。
现有文本到图像扩散模型的概念擦除方法多依赖固定锚点，常导致概念复现和语义侵蚀。我们通过因果追踪揭示擦除对锚点选择的敏感性，定义了更优的‘兄弟排他概念’作为锚点。基于此，提出SELECT（Sibling-Exclusive Evaluation for Contextual Targeting）动态锚点选择框架，采用两阶段评估机制，自动发现最优擦除锚点并识别关键边界锚点以保护相关概念。大量实验表明，SELECT作为通用锚点方案，不仅能高效适配多种擦除框架，且在关键性能指标上持续优于现有基线，单概念锚点挖掘平均仅需4秒。
原文摘要 · Abstract (English)
Existing concept erasure methods for text-to-image diffusion models commonly rely on fixed anchor strategies, which often lead to critical issues such as concept re-emergence and erosion. To address this, we conduct causal tracing to reveal the inherent sensitivity of erasure to anchor selection and define Sibling Exclusive Concepts as a superior class of anchors. Based on this insight, we propose \textbf{SELECT} (Sibling-Exclusive Evaluation for Contextual Targeting), a dynamic anchor selection framework designed to overcome the limitations of fixed anchors. Our framework introduces a novel two-stage evaluation mechanism that automatically discovers optimal anchors for precise erasure while identifying critical boundary anchors to preserve related concepts. Extensive evaluations demonstrate that SELECT, as a universal anchor solution, not only efficiently adapts to multiple erasure frameworks but also consistently outperforms existing baselines across key performance metrics, averaging only 4 seconds for anchor mining of a single concept.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。