{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7SN2QPAVBTWLE3KJHKJEAOACWZ","short_pith_number":"pith:7SN2QPAV","schema_version":"1.0","canonical_sha256":"fc9ba83c150cecb26d493a92403802b654d5ec87cb6e46896fb7f1f510a26504","source":{"kind":"arxiv","id":"2608.10837","version":1},"attestation_state":"computed","paper":{"title":"TACTICL: Task-Aware Compression of Tabular ICL Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Katharina Eggensperger, Matthias Feurer, Mykhailo Koshil","submitted_at":"2026-08-11T12:03:37Z","abstract_excerpt":"The strong performance of foundation models for tabular tasks comes at substantial inference costs. Distilling models into task-specific architectures reduces model size and computational demands but also sacrifices in-context adaptability. Here we introduce TACTICL, an automated task-aware compression framework for tabular in-context learning models that jointly prunes transformer layers and replaces them with lightweight adapters trained on downstream tasks, thus blending in-context with in-weight learning. We study TACTICL on 47 benchmark datasets and show that we can substitute up to 85% o"},"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":"2608.10837","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-11T12:03:37Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ae0f3f1b3af32eb279c7ccee48aa2e1184f36b6968c84ed18a28a071b88f3cd2","abstract_canon_sha256":"1a11c7f588ccc8d4af9ab76bec4ab895fd6388031ad1875918ead0ae621749cf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-12T01:23:27.057144Z","signature_b64":"hK99kIEQzNX6TdUMuqL4LUrhTBNXS8txe1fooA25N6lvsbZuc0iT4JFN5j1gYXP7snQt0KKDN2+ccwShjKziAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc9ba83c150cecb26d493a92403802b654d5ec87cb6e46896fb7f1f510a26504","last_reissued_at":"2026-08-12T01:23:27.054883Z","signature_status":"signed_v1","first_computed_at":"2026-08-12T01:23:27.054883Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TACTICL: Task-Aware Compression of Tabular ICL Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Katharina Eggensperger, Matthias Feurer, Mykhailo Koshil","submitted_at":"2026-08-11T12:03:37Z","abstract_excerpt":"The strong performance of foundation models for tabular tasks comes at substantial inference costs. Distilling models into task-specific architectures reduces model size and computational demands but also sacrifices in-context adaptability. Here we introduce TACTICL, an automated task-aware compression framework for tabular in-context learning models that jointly prunes transformer layers and replaces them with lightweight adapters trained on downstream tasks, thus blending in-context with in-weight learning. We study TACTICL on 47 benchmark datasets and show that we can substitute up to 85% o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.10837","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/2608.10837/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":"2608.10837","created_at":"2026-08-12T01:23:27.058876+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.10837v1","created_at":"2026-08-12T01:23:27.058876+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.10837","created_at":"2026-08-12T01:23:27.058876+00:00"},{"alias_kind":"pith_short_12","alias_value":"7SN2QPAVBTWL","created_at":"2026-08-12T01:23:27.058876+00:00"},{"alias_kind":"pith_short_16","alias_value":"7SN2QPAVBTWLE3KJ","created_at":"2026-08-12T01:23:27.058876+00:00"},{"alias_kind":"pith_short_8","alias_value":"7SN2QPAV","created_at":"2026-08-12T01:23:27.058876+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ","json":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ.json","graph_json":"https://pith.science/api/pith-number/7SN2QPAVBTWLE3KJHKJEAOACWZ/graph.json","events_json":"https://pith.science/api/pith-number/7SN2QPAVBTWLE3KJHKJEAOACWZ/events.json","paper":"https://pith.science/paper/7SN2QPAV"},"agent_actions":{"view_html":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ","download_json":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ.json","view_paper":"https://pith.science/paper/7SN2QPAV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.10837&json=true","fetch_graph":"https://pith.science/api/pith-number/7SN2QPAVBTWLE3KJHKJEAOACWZ/graph.json","fetch_events":"https://pith.science/api/pith-number/7SN2QPAVBTWLE3KJHKJEAOACWZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ/action/storage_attestation","attest_author":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ/action/author_attestation","sign_citation":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ/action/citation_signature","submit_replication":"https://pith.science/pith/7SN2QPAVBTWLE3KJHKJEAOACWZ/action/replication_record"}},"created_at":"2026-08-12T01:23:27.058876+00:00","updated_at":"2026-08-12T01:23:27.058876+00:00"}