Sparse autoencoder features get a ground-truth ruler: seed reproducibility is not concept alignment, and a five-seed consensus reaches 78–98%
Sparse autoencoders (SAEs) are widely used to pull interpretable features from biological foundation-model embeddings, and reproducibility across random seeds has served as the trust test. On embeddings with known concepts, single SAE fits reproduced only 27–46% of their features yet covered most concepts, while a consensus over five fits raised reproducibility to 78–98% at single-fit purity. The consensus covered fewer lineages than single fits, and all SAEs missed weak cell states.