AI RESEARCH
From Token Lists to Graph Motifs: Weisfeiler-Lehman Analysis of Sparse Autoencoder Features
arXiv CS.AI
•
ArXi:2605.06494v1 Announce Type: new Sparse autoencoders (SAEs) have become central to mechanistic interpretability, decomposing transformer activations into monosemantic features. Yet existing analyses characterise features almost exclusively through top-activating token lists or decoder weight vectors, leaving the higher-order co-occurrence structure shared across features largely unexamined. We