{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4U72QVPQ3LXRNVFHFKQA4HXKKZ","short_pith_number":"pith:4U72QVPQ","schema_version":"1.0","canonical_sha256":"e53fa855f0daef16d4a72aa00e1eea566debc71ff3c6ed1cdec52d9a337355a4","source":{"kind":"arxiv","id":"2505.05427","version":1},"attestation_state":"computed","paper":{"title":"Ultra-FineWeb: Efficient Data Filtering and Verification for High-Quality LLM Training Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chaojun Xiao, Guoyang Zeng, Hongya Lyu, Jie Cai, Jie Zhou, Peijun Tang, Xu Han, Yewei Fang, Yudong Wang, Zhiyuan Liu, Zhi Zheng, Zixuan Fu","submitted_at":"2025-05-08T17:15:20Z","abstract_excerpt":"Data quality has become a key factor in enhancing model performance with the rapid development of large language models (LLMs). Model-driven data filtering has increasingly become a primary approach for acquiring high-quality data. However, it still faces two main challenges: (1) the lack of an efficient data verification strategy makes it difficult to provide timely feedback on data quality; and (2) the selection of seed data for training classifiers lacks clear criteria and relies heavily on human expertise, introducing a degree of subjectivity. To address the first challenge, we introduce a"},"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":"2505.05427","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-08T17:15:20Z","cross_cats_sorted":[],"title_canon_sha256":"315fdfcbd506e9c5a7619a484c4acdaed297cf678ad09b9f52494a8e45420325","abstract_canon_sha256":"8ab34fdf85020432ac99c2c9bd85c6a4f403601538626506339be0a64aad4343"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:00:24.316459Z","signature_b64":"yl1oDYCpkPcnFrOOM8xd2Pcet+rSaBpg/11G3F+SmE4aF8G0pG7fjIP7dafkwcpNC5rMaXF7ltOmt36We0WcAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e53fa855f0daef16d4a72aa00e1eea566debc71ff3c6ed1cdec52d9a337355a4","last_reissued_at":"2026-07-05T11:00:24.315881Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:00:24.315881Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Ultra-FineWeb: Efficient Data Filtering and Verification for High-Quality LLM Training Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chaojun Xiao, Guoyang Zeng, Hongya Lyu, Jie Cai, Jie Zhou, Peijun Tang, Xu Han, Yewei Fang, Yudong Wang, Zhiyuan Liu, Zhi Zheng, Zixuan Fu","submitted_at":"2025-05-08T17:15:20Z","abstract_excerpt":"Data quality has become a key factor in enhancing model performance with the rapid development of large language models (LLMs). Model-driven data filtering has increasingly become a primary approach for acquiring high-quality data. However, it still faces two main challenges: (1) the lack of an efficient data verification strategy makes it difficult to provide timely feedback on data quality; and (2) the selection of seed data for training classifiers lacks clear criteria and relies heavily on human expertise, introducing a degree of subjectivity. To address the first challenge, we introduce a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.05427","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/2505.05427/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":"2505.05427","created_at":"2026-07-05T11:00:24.315945+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.05427v1","created_at":"2026-07-05T11:00:24.315945+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.05427","created_at":"2026-07-05T11:00:24.315945+00:00"},{"alias_kind":"pith_short_12","alias_value":"4U72QVPQ3LXR","created_at":"2026-07-05T11:00:24.315945+00:00"},{"alias_kind":"pith_short_16","alias_value":"4U72QVPQ3LXRNVFH","created_at":"2026-07-05T11:00:24.315945+00:00"},{"alias_kind":"pith_short_8","alias_value":"4U72QVPQ","created_at":"2026-07-05T11:00:24.315945+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.21166","citing_title":"The Ratchet Effect in Silico: How Interaction Drives Cumulative Intelligence in Large Language Models","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2510.25741","citing_title":"Scaling Latent Reasoning via Looped Language Models","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ","json":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ.json","graph_json":"https://pith.science/api/pith-number/4U72QVPQ3LXRNVFHFKQA4HXKKZ/graph.json","events_json":"https://pith.science/api/pith-number/4U72QVPQ3LXRNVFHFKQA4HXKKZ/events.json","paper":"https://pith.science/paper/4U72QVPQ"},"agent_actions":{"view_html":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ","download_json":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ.json","view_paper":"https://pith.science/paper/4U72QVPQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.05427&json=true","fetch_graph":"https://pith.science/api/pith-number/4U72QVPQ3LXRNVFHFKQA4HXKKZ/graph.json","fetch_events":"https://pith.science/api/pith-number/4U72QVPQ3LXRNVFHFKQA4HXKKZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ/action/storage_attestation","attest_author":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ/action/author_attestation","sign_citation":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ/action/citation_signature","submit_replication":"https://pith.science/pith/4U72QVPQ3LXRNVFHFKQA4HXKKZ/action/replication_record"}},"created_at":"2026-07-05T11:00:24.315945+00:00","updated_at":"2026-07-05T11:00:24.315945+00:00"}