SOURCE-LINKED INTELLIGENCE

Gaussian Core LoRA: Distribution-Aware Dynamic Adaptation for Broad Concept Erasure

arXiv · AI, language, vision and robotics · article · Sep 1, 2026 · UTC

Concept erasure aims to suppress unsafe, privacy-sensitive, or undesirable generations in text-to-image diffusion models while preserving benign semantics, visual quality, and deployment efficiency. Existing adapter-based methods, such as Low-Rank Adaptation (LoRA), typically freeze the diffusion backbone and learn lightweight parameter updates to steer generation away from target semantics. However, these methods usually assign a static semantic erasure direction to each target concept. This assumption is overly coarse for broad and complex target concepts, since a concept often contains mult

Read original source ↗ Open in workspace

recordType

paper

region

Global