{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TMAXGR4ORIFNTRVWV4KAYQOPB2","short_pith_number":"pith:TMAXGR4O","schema_version":"1.0","canonical_sha256":"9b0173478e8a0ad9c6b6af140c41cf0e9509bd71dcc8f3c352d320002cf46053","source":{"kind":"arxiv","id":"2502.19363","version":3},"attestation_state":"computed","paper":{"title":"DataMan: Data Manager for Pre-training Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dayiheng Liu, Junbo Zhao, Junyang Lin, Kexin Yang, Ru Peng, Yawen Zeng","submitted_at":"2025-02-26T18:01:19Z","abstract_excerpt":"The performance emergence of large language models (LLMs) driven by data scaling laws makes the selection of pre-training data increasingly important. However, existing methods rely on limited heuristics and human intuition, lacking comprehensive and clear guidelines. To address this, we are inspired by ``reverse thinking'' -- prompting LLMs to self-identify which criteria benefit its performance. As its pre-training capabilities are related to perplexity (PPL), we derive 14 quality criteria from the causes of text perplexity anomalies and introduce 15 common application domains to support dom"},"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":"2502.19363","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-26T18:01:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3052739ee0d20eecd8c0d4422dc0dc3f50f68401dc11772a6618f5b81a82da3c","abstract_canon_sha256":"f3020e84ade40c77480dc1317eb592181d9b612e9790afaf9cb3eb75a476ac00"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:46:01.079890Z","signature_b64":"oTVZZ+ajfSO9tFleeDF3ZUhYv2kBurl1lMCJ93xQ3RPpPYsMGk0e6YUSe58Zd/4NI85+K6xpYJdwtfcfTHONDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9b0173478e8a0ad9c6b6af140c41cf0e9509bd71dcc8f3c352d320002cf46053","last_reissued_at":"2026-07-05T10:46:01.079335Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:46:01.079335Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DataMan: Data Manager for Pre-training Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dayiheng Liu, Junbo Zhao, Junyang Lin, Kexin Yang, Ru Peng, Yawen Zeng","submitted_at":"2025-02-26T18:01:19Z","abstract_excerpt":"The performance emergence of large language models (LLMs) driven by data scaling laws makes the selection of pre-training data increasingly important. However, existing methods rely on limited heuristics and human intuition, lacking comprehensive and clear guidelines. To address this, we are inspired by ``reverse thinking'' -- prompting LLMs to self-identify which criteria benefit its performance. As its pre-training capabilities are related to perplexity (PPL), we derive 14 quality criteria from the causes of text perplexity anomalies and introduce 15 common application domains to support dom"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.19363","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/2502.19363/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":"2502.19363","created_at":"2026-07-05T10:46:01.079398+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.19363v3","created_at":"2026-07-05T10:46:01.079398+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.19363","created_at":"2026-07-05T10:46:01.079398+00:00"},{"alias_kind":"pith_short_12","alias_value":"TMAXGR4ORIFN","created_at":"2026-07-05T10:46:01.079398+00:00"},{"alias_kind":"pith_short_16","alias_value":"TMAXGR4ORIFNTRVW","created_at":"2026-07-05T10:46:01.079398+00:00"},{"alias_kind":"pith_short_8","alias_value":"TMAXGR4O","created_at":"2026-07-05T10:46:01.079398+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.05267","citing_title":"Bridging Generation and Training: A Systematic Review of Quality Issues in LLMs for Code","ref_index":97,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2","json":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2.json","graph_json":"https://pith.science/api/pith-number/TMAXGR4ORIFNTRVWV4KAYQOPB2/graph.json","events_json":"https://pith.science/api/pith-number/TMAXGR4ORIFNTRVWV4KAYQOPB2/events.json","paper":"https://pith.science/paper/TMAXGR4O"},"agent_actions":{"view_html":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2","download_json":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2.json","view_paper":"https://pith.science/paper/TMAXGR4O","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.19363&json=true","fetch_graph":"https://pith.science/api/pith-number/TMAXGR4ORIFNTRVWV4KAYQOPB2/graph.json","fetch_events":"https://pith.science/api/pith-number/TMAXGR4ORIFNTRVWV4KAYQOPB2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2/action/storage_attestation","attest_author":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2/action/author_attestation","sign_citation":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2/action/citation_signature","submit_replication":"https://pith.science/pith/TMAXGR4ORIFNTRVWV4KAYQOPB2/action/replication_record"}},"created_at":"2026-07-05T10:46:01.079398+00:00","updated_at":"2026-07-05T10:46:01.079398+00:00"}