AI RESEARCH
SALVE: Sparse Autoencoder-Latent Vector Editing for Mechanistic Control of Neural Networks
arXiv CS.LG
•
ArXi:2512.15938v2 Announce Type: replace 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.