Recent deep learning models for tabular data currently compete with the traditional ML models based on decision trees (GBDT). Unlike GBDT, deep models can additionally benefit from pretraining, which is a workhorse of DL for vision and NLP. For tabular problems, several pretraining methods were proposed, but it is not entirely clear if pretraining provides consistent noticeable improvements and what method should be used, since the methods are often not compared to each other or comparison is limited to the simplest MLP architectures. In this work, we aim to identify the best practices to prResearch goal: How do different adversarial perturbation techniques during pretraining (e.g., FGSM, PGD) influence the alignment and fairness of tabular foundation models, as evaluated by metrics like demographic parity on tabular fairness benchmarks like TabFair?Autonomous synthesis report generated by SOVEREIGN Research Kernel. Tribunal consensus score: 7.9/10.