{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NZAABDPB6Y4E7IUMXJIFU5UYLJ","short_pith_number":"pith:NZAABDPB","schema_version":"1.0","canonical_sha256":"6e40008de1f6384fa28cba505a76985a55102b51aae81fe956664a82d1e45dc2","source":{"kind":"arxiv","id":"2501.08496","version":3},"attestation_state":"computed","paper":{"title":"Quantifying the Importance of Data Alignment in Downstream Model Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.PL"],"primary_cat":"cs.CL","authors_text":"Aryan Sahai, Brando Miranda, Elyas Obbad, Krrish Chawla, Mario DePavia, Sanmi Koyejo, Sudharsan Sundar","submitted_at":"2025-01-14T23:59:23Z","abstract_excerpt":"Contrary to the conventional emphasis on dataset size, we explore the role of data alignment -- an often overlooked aspect of data quality -- in training capable Large Language Models (LLMs). To do so, we use the Task2Vec-based alignment coefficient, a quantitative measure of the similarity between two datasets, to quantify the impact of alignment between training data and evaluation data on downstream performance. In particular, we conduct controlled \\textit{interventional} experiments for two settings: 1. the impact of increased alignment coefficients between various pre-training (pt) agains"},"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":"2501.08496","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-14T23:59:23Z","cross_cats_sorted":["cs.AI","cs.LG","cs.PL"],"title_canon_sha256":"f0ae926ea0e84c745482a9cd7a7839b790c58c8d3ff7ec81316b8882a9ff0e81","abstract_canon_sha256":"9a554c7b894dc4dcf82e31c96b5f36d658620c811829f3f8890fba286366a7a5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:31:04.329222Z","signature_b64":"SBDTYYQ+llCHTAf4YSps2RL2ZFQcSZaO885LwMAgp0vhcI6lLj0SdXIOci/h6Udbr8fSIHyb6FQnSXAYIF+LDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6e40008de1f6384fa28cba505a76985a55102b51aae81fe956664a82d1e45dc2","last_reissued_at":"2026-07-05T11:31:04.328695Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:31:04.328695Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Quantifying the Importance of Data Alignment in Downstream Model Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.PL"],"primary_cat":"cs.CL","authors_text":"Aryan Sahai, Brando Miranda, Elyas Obbad, Krrish Chawla, Mario DePavia, Sanmi Koyejo, Sudharsan Sundar","submitted_at":"2025-01-14T23:59:23Z","abstract_excerpt":"Contrary to the conventional emphasis on dataset size, we explore the role of data alignment -- an often overlooked aspect of data quality -- in training capable Large Language Models (LLMs). To do so, we use the Task2Vec-based alignment coefficient, a quantitative measure of the similarity between two datasets, to quantify the impact of alignment between training data and evaluation data on downstream performance. In particular, we conduct controlled \\textit{interventional} experiments for two settings: 1. the impact of increased alignment coefficients between various pre-training (pt) agains"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.08496","kind":"arxiv","version":3},"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/2501.08496/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":"2501.08496","created_at":"2026-07-05T11:31:04.328748+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.08496v3","created_at":"2026-07-05T11:31:04.328748+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.08496","created_at":"2026-07-05T11:31:04.328748+00:00"},{"alias_kind":"pith_short_12","alias_value":"NZAABDPB6Y4E","created_at":"2026-07-05T11:31:04.328748+00:00"},{"alias_kind":"pith_short_16","alias_value":"NZAABDPB6Y4E7IUM","created_at":"2026-07-05T11:31:04.328748+00:00"},{"alias_kind":"pith_short_8","alias_value":"NZAABDPB","created_at":"2026-07-05T11:31:04.328748+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.19667","citing_title":"Survey of Specialized Large Language Model","ref_index":8,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ","json":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ.json","graph_json":"https://pith.science/api/pith-number/NZAABDPB6Y4E7IUMXJIFU5UYLJ/graph.json","events_json":"https://pith.science/api/pith-number/NZAABDPB6Y4E7IUMXJIFU5UYLJ/events.json","paper":"https://pith.science/paper/NZAABDPB"},"agent_actions":{"view_html":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ","download_json":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ.json","view_paper":"https://pith.science/paper/NZAABDPB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.08496&json=true","fetch_graph":"https://pith.science/api/pith-number/NZAABDPB6Y4E7IUMXJIFU5UYLJ/graph.json","fetch_events":"https://pith.science/api/pith-number/NZAABDPB6Y4E7IUMXJIFU5UYLJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ/action/storage_attestation","attest_author":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ/action/author_attestation","sign_citation":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ/action/citation_signature","submit_replication":"https://pith.science/pith/NZAABDPB6Y4E7IUMXJIFU5UYLJ/action/replication_record"}},"created_at":"2026-07-05T11:31:04.328748+00:00","updated_at":"2026-07-05T11:31:04.328748+00:00"}