Information Routing in Atomistic Foundation Models: How Task Alignment and Equivariance Shape Linear Disentanglement
Joshua Steier · 2026
Read Information Routing in Atomistic Foundation Models: How Task Alignment and Equivariance Shape Linear Disentanglement by Joshua Steier (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-03155.txt.
Description
Joshua Steier asks what shapes the inner organization of models that predict molecular properties. Some models keep shape information cleanly separable from composition, others tangle the two. His 2026 paper introduces a probe method that strips out the composition signal and measures how much geometric signal a simple linear reader can still recover.
Across ten models from five families, recoverable geometry varied more than sixfold. The training target dominated: models trained on an energy-gap property kept far more geometry than energy-trained ones, whatever their architecture. Richer training data partly compensated for a mismatched target. Inside one model family, different symmetry channels specialized, vectors carrying dipole information and scalars carrying gap information. Nonlinear probes, he warns, can fake separation that is not there.
Catalog record
| CREATOR | Joshua Steier |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 82801 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.03155