{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:77ULOZN66U2XGCDFVF24BJYUSI","short_pith_number":"pith:77ULOZN6","schema_version":"1.0","canonical_sha256":"ffe8b765bef535730865a975c0a7149212d4891bb66750c3b284f7dbbf7a72f0","source":{"kind":"arxiv","id":"2507.17066","version":1},"attestation_state":"computed","paper":{"title":"Risk In Context: Benchmarking Privacy Leakage of Foundation Models in Synthetic Tabular Data Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Guang Cheng, Jessup Byun, Joshua Ward, Xiaofeng Lin","submitted_at":"2025-07-22T22:59:08Z","abstract_excerpt":"Synthetic tabular data is essential for machine learning workflows, especially for expanding small or imbalanced datasets and enabling privacy-preserving data sharing. However, state-of-the-art generative models (GANs, VAEs, diffusion models) rely on large datasets with thousands of examples. In low-data settings, often the primary motivation for synthetic data, these models can overfit, leak sensitive records, and require frequent retraining. Recent work uses large pre-trained transformers to generate rows via in-context learning (ICL), which needs only a few seed examples and no parameter up"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2507.17066","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-22T22:59:08Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"8b82b73ba25cffe5279bc3b1a643f741d8252f0de1ef568aa81c228d4451ca86","abstract_canon_sha256":"c6d47f2898311aa78d1da088780be9d981c9a63b4219163a0ce98028341cdbe1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:42:01.050556Z","signature_b64":"6MxpdZasWjY10KUwrBw46TRSre4njGMRHpNk6eCJEWlE7qgV5SbAi3GyKTQakg68hunRsVGU6lYFxXxsgD03Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ffe8b765bef535730865a975c0a7149212d4891bb66750c3b284f7dbbf7a72f0","last_reissued_at":"2026-07-05T11:42:01.050130Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:42:01.050130Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Risk In Context: Benchmarking Privacy Leakage of Foundation Models in Synthetic Tabular Data Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Guang Cheng, Jessup Byun, Joshua Ward, Xiaofeng Lin","submitted_at":"2025-07-22T22:59:08Z","abstract_excerpt":"Synthetic tabular data is essential for machine learning workflows, especially for expanding small or imbalanced datasets and enabling privacy-preserving data sharing. However, state-of-the-art generative models (GANs, VAEs, diffusion models) rely on large datasets with thousands of examples. In low-data settings, often the primary motivation for synthetic data, these models can overfit, leak sensitive records, and require frequent retraining. Recent work uses large pre-trained transformers to generate rows via in-context learning (ICL), which needs only a few seed examples and no parameter up"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.17066","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2507.17066/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2507.17066","created_at":"2026-07-05T11:42:01.050188+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.17066v1","created_at":"2026-07-05T11:42:01.050188+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.17066","created_at":"2026-07-05T11:42:01.050188+00:00"},{"alias_kind":"pith_short_12","alias_value":"77ULOZN66U2X","created_at":"2026-07-05T11:42:01.050188+00:00"},{"alias_kind":"pith_short_16","alias_value":"77ULOZN66U2XGCDF","created_at":"2026-07-05T11:42:01.050188+00:00"},{"alias_kind":"pith_short_8","alias_value":"77ULOZN6","created_at":"2026-07-05T11:42:01.050188+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26021","citing_title":"Privacy Vulnerabilities of Attention Layers in Tabular Foundation Models and Protection of High-Risk Queries","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2606.30410","citing_title":"Beyond IID: How General Are Tabular Foundation Models, Really?","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2512.08875","citing_title":"When Tables Leak: Attacking String Memorization in LLM-Based Tabular Data Generation","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI","json":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI.json","graph_json":"https://pith.science/api/pith-number/77ULOZN66U2XGCDFVF24BJYUSI/graph.json","events_json":"https://pith.science/api/pith-number/77ULOZN66U2XGCDFVF24BJYUSI/events.json","paper":"https://pith.science/paper/77ULOZN6"},"agent_actions":{"view_html":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI","download_json":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI.json","view_paper":"https://pith.science/paper/77ULOZN6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.17066&json=true","fetch_graph":"https://pith.science/api/pith-number/77ULOZN66U2XGCDFVF24BJYUSI/graph.json","fetch_events":"https://pith.science/api/pith-number/77ULOZN66U2XGCDFVF24BJYUSI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI/action/storage_attestation","attest_author":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI/action/author_attestation","sign_citation":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI/action/citation_signature","submit_replication":"https://pith.science/pith/77ULOZN66U2XGCDFVF24BJYUSI/action/replication_record"}},"created_at":"2026-07-05T11:42:01.050188+00:00","updated_at":"2026-07-05T11:42:01.050188+00:00"}