When Do Learned Blocks Preserve Causal Meaning? Block Identifiability and Constructive Causal Abstractions
arXiv:2610.18155v2 Announce Type: replace-cross Abstract: Standard identifiability theory assumes one-dimensional latent factors, but learned representations can encode latent information across multidimensional blocks whose internal coordinates are not uniquely determined. When such blocks are used to represent and intervene on higher-level concepts, it is unclear whether their causal meaning survives the transformations left unresolved by block-identifiability. We connect block-identifiabilit…