TabAbax 技术报告：聚焦合成先验与编码器的表格基础模型
TabAbax: A Tabular Foundation Model Technical Report
一家研究团队在 TabICLv2 基础上做了个表格基础模型 TabAbax，回归任务在 TabArena 上比原版高了 300 多 Elo，还用消融实验证明合成先验才是关键。
TabAbax 是在 TabICLv2 基础上构建的表格基础模型（TFM），改动集中在合成预训练先验和列-行编码器两个上游组件。在 142 数据集的 BeyondArena 套件上拿到 1,238 Elo，位列 41 个方法中的第 10；在 TabArena 51 数据集总榜上 1,676 Elo，比 TabICLv2 高 115 分。回归任务提升最明显：TabArena 回归基准 2,024 Elo，超过 TabICLv2 336 分、超过 TabFM 21 分。消融实验显示改进后的先验单独带来回归 +228 Elo，证明先验质量是回归性能的主导因素。
TabAbax: A Tabular Foundation Model Technical Report
Tabular foundation models (TFMs) enable in-context prediction on unseen tabular datasets without task-specific training. Mechanistic work has shown that increasing the depth of the ICL Transformer yields weak gains, leaving the main headroom for gains in the synthetic pretraining prior and column-row encoder upstream of the ICL backbone. We introduce TabAbax, a TFM built on TabICLv2 that focuses all modifications on these two upstream components. On the prior side, we further improve the quality and complexity of the synthetic pretraining prior. On the encoder side, a two-pass column-row stack strengthens cross-feature interactions before row compression, using learned Fourier scalar encoding from TabFM and Causilo. Evaluations on BeyondArena and TabArena show strong regression-centric gains: TabAbax reaches 1,238 Elo on the 142-dataset BeyondArena suite (10th of 41 evaluated methods), 1,676 Elo on TabArena's overall 51-dataset suite (11th of 99 methods, +115 over TabICLv2), and 2,024 Elo on TabArena's regression benchmark (7th of 99 methods, +336 over TabICLv2, +21 over TabFM). A prior-only ablation confirms the attribution: the improved prior alone yields +228 Elo for regression, demonstrating that prior quality is the dominant knob for improving regression performance.