Boltz is a Strong Baseline for Atom-level Representation Learning
Hyosoon Jang, Hyunjin Seo, Yunhui Jang, Seonghyun Park, Sungsoo Ahn · 2026
Read Boltz is a Strong Baseline for Atom-level Representation Learning by Hyosoon Jang, Hyunjin Seo, Yunhui Jang, Seonghyun Park, Sungsoo Ahn (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-2602-13249.txt.
Description
Small-molecule models are usually trained on lone molecules, missing the rich clues that come from seeing a molecule bound to a protein. This study asks whether a protein-ligand co-folding model can supply those clues instead, testing the Boltz2 model by moving its atom-level ligand readings onto stand-alone molecule tasks. The transfer is checked through probing and distillation, two ways of reusing what a trained model already knows.
Boltz2 matched or beat existing models on the ADMET benchmark, sped up molecule generation, and used samples more efficiently when guiding ligand design. Its readings also added value beyond standard training signals such as 3D shapes, lab assay labels, and quantum chemistry traits. Hyosoon Jang leads the work with six co-authors, and the authors present co-folding as a strong base for molecule models.
Catalog record
| CREATOR | Hyosoon Jang, Hyunjin Seo, Yunhui Jang, Seonghyun Park, Sungsoo Ahn |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-13249