A Single Neuron Works: Precise Concept Erasure in Text-to-Image Diffusion Models
This addresses safety risks in text-to-image generation, offering a precise solution for concept erasure, though it is incremental as it builds on existing methods like sparse autoencoders.
The paper tackles the problem of precisely erasing harmful concepts in text-to-image diffusion models while minimizing image quality degradation, achieving state-of-the-art results by manipulating only a single neuron.
Text-to-image models exhibit remarkable capabilities in image generation. However, they also pose safety risks of generating harmful content. A key challenge of existing concept erasure methods is the precise removal of target concepts while minimizing degradation of image quality. In this paper, we propose Single Neuron-based Concept Erasure (SNCE), a novel approach that can precisely prevent harmful content generation by manipulating only a single neuron. Specifically, we train a Sparse Autoencoder (SAE) to map text embeddings into a sparse, disentangled latent space, where individual neurons align tightly with atomic semantic concepts. To accurately locate neurons responsible for harmful concepts, we design a novel neuron identification method based on the modulated frequency scoring of activation patterns. By suppressing activations of the harmful concept-specific neuron, SNCE achieves surgical precision in concept erasure with minimal disruption to image quality. Experiments on various benchmarks demonstrate that SNCE achieves state-of-the-art results in target concept erasure, while preserving the model's generation capabilities for non-target concepts. Additionally, our method exhibits strong robustness against adversarial attacks, significantly outperforming existing methods.