{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Z4REOJCG3J5PCNSR74CXD4LJQJ","short_pith_number":"pith:Z4REOJCG","schema_version":"1.0","canonical_sha256":"cf22472446da7af13651ff0571f169826ec8fe29d34314ef723925eb9d5321ad","source":{"kind":"arxiv","id":"2506.10952","version":1},"attestation_state":"computed","paper":{"title":"Domain2Vec: Vectorizing Datasets to Find the Optimal Data Mixture without Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Howe Tissue, Lu Wang, Mozhi Zhang, Xipeng Qiu","submitted_at":"2025-06-12T17:53:51Z","abstract_excerpt":"We introduce~\\textsc{Domain2Vec}, a novel approach that decomposes any dataset into a linear combination of several \\emph{meta-domains}, a new concept designed to capture the key underlying features of datasets. \\textsc{Domain2Vec} maintains a vocabulary of meta-domains and uses a classifier to decompose any given dataset into a domain vector that corresponds to a distribution over this vocabulary. These domain vectors enable the identification of the optimal data mixture for language model (LM) pretraining in a training-free manner under the \\emph{\\textbf{D}istribution \\textbf{A}lignment \\tex"},"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":"2506.10952","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-12T17:53:51Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"b32e9a84213e1810f111dd3998275fbfdd68b15f530bbc1e553367860a594d35","abstract_canon_sha256":"fd331797bb474d735499fde9618e84f80ca209f692e5bf826be6276863dbf71f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:20:40.837113Z","signature_b64":"oJ1MiNAEN0r3yM/MSP1fxdCCQMnpud1swnlSA9WUNr46gM36VX2hz9tC6B3Y3jchVE9SpQ8oQ2hVUme+CHlpBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cf22472446da7af13651ff0571f169826ec8fe29d34314ef723925eb9d5321ad","last_reissued_at":"2026-07-05T11:20:40.836680Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:20:40.836680Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Domain2Vec: Vectorizing Datasets to Find the Optimal Data Mixture without Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Howe Tissue, Lu Wang, Mozhi Zhang, Xipeng Qiu","submitted_at":"2025-06-12T17:53:51Z","abstract_excerpt":"We introduce~\\textsc{Domain2Vec}, a novel approach that decomposes any dataset into a linear combination of several \\emph{meta-domains}, a new concept designed to capture the key underlying features of datasets. \\textsc{Domain2Vec} maintains a vocabulary of meta-domains and uses a classifier to decompose any given dataset into a domain vector that corresponds to a distribution over this vocabulary. These domain vectors enable the identification of the optimal data mixture for language model (LM) pretraining in a training-free manner under the \\emph{\\textbf{D}istribution \\textbf{A}lignment \\tex"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.10952","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/2506.10952/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":"2506.10952","created_at":"2026-07-05T11:20:40.836743+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.10952v1","created_at":"2026-07-05T11:20:40.836743+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.10952","created_at":"2026-07-05T11:20:40.836743+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z4REOJCG3J5P","created_at":"2026-07-05T11:20:40.836743+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z4REOJCG3J5PCNSR","created_at":"2026-07-05T11:20:40.836743+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z4REOJCG","created_at":"2026-07-05T11:20:40.836743+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.16380","citing_title":"Data Mixing for Large Language Models Pretraining: A Survey and Outlook","ref_index":65,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ","json":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ.json","graph_json":"https://pith.science/api/pith-number/Z4REOJCG3J5PCNSR74CXD4LJQJ/graph.json","events_json":"https://pith.science/api/pith-number/Z4REOJCG3J5PCNSR74CXD4LJQJ/events.json","paper":"https://pith.science/paper/Z4REOJCG"},"agent_actions":{"view_html":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ","download_json":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ.json","view_paper":"https://pith.science/paper/Z4REOJCG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.10952&json=true","fetch_graph":"https://pith.science/api/pith-number/Z4REOJCG3J5PCNSR74CXD4LJQJ/graph.json","fetch_events":"https://pith.science/api/pith-number/Z4REOJCG3J5PCNSR74CXD4LJQJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ/action/storage_attestation","attest_author":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ/action/author_attestation","sign_citation":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ/action/citation_signature","submit_replication":"https://pith.science/pith/Z4REOJCG3J5PCNSR74CXD4LJQJ/action/replication_record"}},"created_at":"2026-07-05T11:20:40.836743+00:00","updated_at":"2026-07-05T11:20:40.836743+00:00"}