{"ID":6538288,"CreatedAt":"2026-07-14T02:54:43.516908796Z","UpdatedAt":"2026-07-15T03:28:55.185153975Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.11007","arxiv_id":"2607.11007","title":"TabPFN beyond Tabular Data: Calibration and Accuracy on Multimodal Embeddings","abstract":"Few-shot multimodal classification commonly attaches a lightweight head, such as $k$-nearest neighbors, logistic regression, or a linear SVM, to a frozen pretrained encoder. Although computationally efficient, these heads can produce poorly calibrated confidence scores, limiting their reliability in calibration-sensitive applications. We evaluate TabPFN as a plug-and-play, zero-gradient classification head for frozen image, text, and audio encoders. Across 22{,}820 evaluation episodes spanning 14 datasets, 11 encoders, and three modalities, TabPFN achieves the best mean rank among nine classification heads on both negative log-likelihood (NLL) and expected calibration error (ECE). At a representative setting, it reduces NLL by 48--62\\% and ECE by 2.1--5.3$\\times$ relative to the average of the eight baselines while matching or exceeding their average accuracy. Its accuracy advantage is conditional, concentrating at moderate-to-high shot counts and low-to-moderate feature dimensions ($k \\ge 50$, $d \\le 32$), and diminishing when labeled data are scarce, feature dimensions are high, or competing methods approach ceiling accuracy. In targeted backbone-adaptation experiments, replacing the trained linear head with TabPFN substantially improves calibration while preserving competitive accuracy. These results provide empirical guidance for using TabPFN as a training-free head in calibration-sensitive multimodal classification. To support transparency and reproducibility, we publicly release the source code, experiment configurations, and evaluation scripts in our GitHub repository: https://github.com/Jingxiang-Zhang/tabpfn-multimodal-embeddings.","short_abstract":"Few-shot multimodal classification commonly attaches a lightweight head, such as $k$-nearest neighbors, logistic regression, or a linear SVM, to a frozen pretrained encoder. Although computationally efficient, these heads can produce poorly calibrated confidence scores, limiting their reliability in calibration-sensiti...","url_abs":"https://arxiv.org/abs/2607.11007","url_pdf":"https://arxiv.org/pdf/2607.11007v1","authors":"[\"Jingxiang Zhang\",\"Lujia Zhong\",\"Zijie Zhu\",\"Shuo Huang\",\"Yuang Xu\"]","published":"2026-07-13T02:11:27Z","proceeding":"cs.LG","tasks":"[\"cs.LG\"]","methods":"[]","has_code":false,"code_links":[{"ID":614232,"CreatedAt":"2026-07-14T02:54:43.516908796Z","UpdatedAt":"2026-07-14T02:54:43.516908796Z","DeletedAt":null,"paper_id":6538288,"paper_url":"https://arxiv.org/abs/2607.11007","paper_title":"TabPFN beyond Tabular Data: Calibration and Accuracy on Multimodal Embeddings","repo_url":"https://github.com/Jingxiang-Zhang/tabpfn-multimodal-embeddings","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}