2 papers
cs.LG2026
Behavioral Compositionality of Sparse Autoencoder Representations
Haotian Ma, Ziye Mei, Frederic Sala
Sparse autoencoders (SAEs) represent model activations with sparse, linearly decoded features, but the downstream effects of those features need not compose additively. We introduc…
cs.LG2026
Representation Redistribution in Sparse Autoencoders
Ziye Mei, Haotian Ma
Sparse autoencoders (SAEs) decompose language model activations into sparse features, but prompt transformations designed to preserve the underlying request can redistribute activi…