Tabular Language Models (TLMs) have been claimed to achieve strong generalization for tabular prediction. We conduct a systematic re-evaluation of Tabula-8B as a representative TLM, utilizing 165 datasets from the UniPredict benchmark. Our investigation reveals three findings. First, binary and categorical classification achieve near-zero median lift over majority-class baselines and strong aggregate performance is driven entirely by quartile classification tasks. Second, top-performing datasets exhibit pervasive contamination, including complete train-test overlap and task-level leakage that evades standard deduplication. Third, instruction-tuning without tabular exposure recovers 92.2% of standard classification performance and on quartile classification, format familiarity closes 71.3% of the gap with the residual attributable to contaminated datasets. These findings suggest claimed generalization likely reflects evaluation artifacts rather than learned tabular reasoning. We conclude with recommendations for strengthening TLM evaluation.
Understanding the Surprising Generalization Properties of Tabular Foundation Models
Polytechnique Montréal（蒙特利尔理工学院）；Mila – Quebec AI Institute（米拉-魁北克人工智能研究所）；Chandar Research Lab（钱达尔研究实验室）；University of Toronto（多伦多大学）；Layer 6 AI（第六层人工智能公司）；Prior Labs（普里奥实验室）；ELLIS Institute Tübingen（埃利斯研究所图宾根分部）；University of Freiburg（弗赖堡大学）；Cohere（科here公司）
Generalization Can Emerge in Tabular Foundation Models From a Single Table
University of Toronto（多伦多大学）；Polytechnique Montréal（蒙特利尔大学）；Mila – Quebec AI Institute（魁北克人工智能研究所）；Layer 6 AI（Layer 6 AI公司）；Prior Labs（Prior实验室）；ELLIS Institute（ELLIS研究所）；Tübingen University of Freiburg（弗赖堡大学）
TabularQGAN: A quantum generative model for tabular data synthesis
SAP SE（SAP公司）；BASF Digital Solutions（巴斯夫数字解决方案）；QUTAC Quantum Technology and Application Consortium（QUTAC量子技术与应用联盟）；Technical University of Munich CIT（慕尼黑技术大学CIT）；Ludwig-Maximilians-Universität München（慕尼黑路德维希-马克西米利安大学）；Merck KGaA（默克公司）
Few-shot Cross-country Generalization of Tabular Machine Learning and Foundation Models for Childhood Anemia Prediction under Distribution Shift
Department of Mathematics, Rhodes University, South Africa（数学系，罗德斯大学，南非）；National Institute for Theoretical and computational Sciences (NITheCS), Stellenbosch, 7600, South Africa（理论与计算科学国家研究所（NITheCS），斯泰伦博斯，7600，南非）；Interdisciplinary Research Program in Public Health, University of Burundi, Burundi（公共卫生跨学科研究计划，布恩迪大学，布恩迪）；Universite Clermont Auvergne, Clermont Auvergne INP, CNRS, Institut Pascal, Clermont–Ferrand, France（克莱蒙特-奥弗涅大学，克莱蒙特-奥弗涅INP，CNRS，帕西尔研究所，克莱蒙特-费尔南，法国）；Department of International Public Health, Liverpool School of Tropical Medicine, Liverpool, UK（国际公共卫生系，利物浦热带医学学校，利物浦，英国）；College of Health Solutions, Arizona State University, Phoenix, USA（健康解决方案学院，亚利桑那州立大学，凤凰城，美国）；Department of Statistics, University of Pretoria, Pretoria, South Africa（统计系，普里特oria大学，普里特oria，南非）
TSPFN: A Temporal Tabular Foundation Model for Physiological Time Series Classification
Sorbonne Université（索邦大学）；CNRS（法国国家科学研究中心）；ISIR（巴黎索邦大学机器人与智能系统研究所）；MLIA（机器学习与人工智能实验室）；Inria（法国国家信息与自动化研究所）；I3S（信息科学与系统实验室）；Institut Universitaire de France (IUF)（法国大学研究院）；INSA-Lyon（里昂国立应用科学学院）；Université Claude Bernard Lyon 1（里昂第一大学）；Inserm（法国国家健康与医学研究院）；CREATIS（图像、信号与智能系统研究中心）
DenoGrad: A Gradient-Based Framework for Data Refinement in Tabular and Time-Series Learning
Department of Computer Science and Artificial Intelligence, University of Granada（格拉纳达大学计算机科学与人工智能系）；Andalusian Institute of Data Science and Computational Intelligence (DaSCI), University of Granada（安达卢西亚数据科学与计算智能研究所（DaSCI），格拉纳达大学）
Learning from Anonymized and Incomplete Tabular Data
Leipzig University \& ScaDS.AI Leipzig Germany；The Australian National University Canberra Australia；Leipzig University \& ScaDS.AI；The Australian National University
