Compositional probe decomposition (CPD)

methods / Direction Extraction / Compositional probe decomposition (CPD)

Techniqueadvanced

Linearly projects out the compositional (atom-count/formula-level) signal from a representation before measuring how much genuinely geometric/structural information remains via a Ridge probe, isolating structure-specific information from a confounding compositional shortcut.

Used in (1 observation)

Atomistic models route geometric information by task and symmetry typemeasured

structure: Linear Subspace · models: PaiNN (polarizable atom interaction neural network), MACE (higher-order equivariant message-passing potential), ViSNet · paper: Information Routing in Atomistic Foundation Models: How Task Alignment and Equivariance Shape Linear Disentanglement