Real-Time Explanations for Tabular Foundation Models
Luan Borges Teodoro Reis Sena, Francisco Galuppo Azevedo · 2026
Read Real-Time Explanations for Tabular Foundation Models by Luan Borges Teodoro Reis Sena, Francisco Galuppo Azevedo (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-2603-29946.txt.
Description
Sena and Galuppo Azevedo present ShapPFN, a tabular model that explains itself in real time. Tabular foundation models predict well, but tools like SHAP explain too slowly for interactive work. ShapPFN bakes Shapley value regression into its design. One forward pass then yields both a prediction and its explanation. On standard benchmarks, accuracy stays competitive. Explanation quality is high, with R squared at 0.96. It runs over a thousand times faster than KernelSHAP.
Catalog record
| CREATOR | Luan Borges Teodoro Reis Sena, Francisco Galuppo Azevedo |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 20328 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.29946