用稀疏编码发现并操控神经网络特征，让模型行为可解释可调控。
SALVE: Sparse Autoencoder-Latent Vector Editing for Mechanistic Control of Neural Networks
- 通过无监督稀疏编码学习模型原生特征基，自动提取关键特征。
- 实现对分类特征和跨类特征的精准永久修改，提升模型可控性。
- 提出关键抑制阈值，支持细粒度鲁棒性分析，适合可解释AI研究者。
深度神经网络性能优异但难以解释与控制。本文提出SALVE（稀疏自编码器-潜在向量编辑）框架，实现特征发现、验证与控制的统一。利用ℓ₁正则化自编码器，在无监督条件下学习稀疏且模型原生的特征基。通过Grad-FAM方法（特征级显著性映射），将潜在特征与输入数据视觉对齐以验证其意义。基于自编码器结构，实现精确且永久的权重空间干预，支持对类别定义特征与跨类别特征的连续调节。进一步推导出关键抑制阈值α_crit，量化每类对主导特征的依赖程度，支持细粒度鲁棒性诊断。在卷积网络（ResNet-18）与Transformer（ViT-B/16）上验证，均实现一致且可解释的模型行为调控。本工作为将特征发现转化为可操作的模型编辑提供系统方法，推动透明可控AI的发展。
原文摘要 · Abstract (English)
Deep neural networks achieve impressive performance but remain difficult to interpret and control. We present SALVE (Sparse Autoencoder-Latent Vector Editing), a unified "discover, validate, and control" framework that bridges mechanistic interpretability and model editing. Using an $\ell_1$-regularized autoencoder, we learn a sparse, model-native feature basis without supervision. We validate these features with Grad-FAM, a feature-level saliency mapping method that visually grounds latent features in input data. Leveraging the autoencoder's structure, we perform precise and permanent weight-space interventions, enabling continuous modulation of both class-defining and cross-class features. We further derive a critical suppression threshold, $α_{crit}$, quantifying each class's reliance on its dominant feature, supporting fine-grained robustness diagnostics. Our approach is validated on both convolutional (ResNet-18) and transformer-based (ViT-B/16) models, demonstrating consistent, interpretable control over their behavior. This work contributes a principled methodology for turning feature discovery into actionable model edits, advancing the development of transparent and controllable AI systems.
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。