{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:EB6HDFZLFDIAXPE4FZV6PTOP4Q","short_pith_number":"pith:EB6HDFZL","schema_version":"1.0","canonical_sha256":"207c71972b28d00bbc9c2e6be7cdcfe419501ba4a5ef6225c2e9a0ac2a18d323","source":{"kind":"arxiv","id":"2605.25698","version":1},"attestation_state":"computed","paper":{"title":"How Should LLMs Consume High-Quality Data? Optimal Data Scheduling via Quality-Aware Functional Scaling Laws","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiawei Fu, Shizhe Wu, Xiaoqing Liu, Xili Wang, Zhitao Zhu","submitted_at":"2026-05-25T10:56:45Z","abstract_excerpt":"High-quality data is scarce in large language model (LLM) training, yet how to schedule its use jointly with training dynamics lacks theoretical guidance. We extend functional scaling laws by incorporating a data-quality dimension, and solve the joint data-quality and batch-size scheduling problem in asymptotic closed form. The solution reveals two regimes and a dual role of high-quality data. In the noise-limited regime, high-quality data should be used as a signal amplifier: lowering the batch size converts cleaner data into more signal without amplifying noise. In the signal-limited regime,"},"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":"2605.25698","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T10:56:45Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c5eaeda72321e8ef2ae354aa23b70f8c1d835a23b9ce81319e5016db7c4d645c","abstract_canon_sha256":"d64ca422ca91c0cfed8032c35334dc558451f8699f9a16bdb7fd93dad62b453e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:04:51.107924Z","signature_b64":"KmuZ8oLDiOGQCK2J6bWif31aqeiN70xr/wiC+wzznzEB5vFzKt7p9zzG8aWRHhqMeN7hhUaXi36YHrsViR4iAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"207c71972b28d00bbc9c2e6be7cdcfe419501ba4a5ef6225c2e9a0ac2a18d323","last_reissued_at":"2026-05-26T02:04:51.107202Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:04:51.107202Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Should LLMs Consume High-Quality Data? Optimal Data Scheduling via Quality-Aware Functional Scaling Laws","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiawei Fu, Shizhe Wu, Xiaoqing Liu, Xili Wang, Zhitao Zhu","submitted_at":"2026-05-25T10:56:45Z","abstract_excerpt":"High-quality data is scarce in large language model (LLM) training, yet how to schedule its use jointly with training dynamics lacks theoretical guidance. We extend functional scaling laws by incorporating a data-quality dimension, and solve the joint data-quality and batch-size scheduling problem in asymptotic closed form. The solution reveals two regimes and a dual role of high-quality data. In the noise-limited regime, high-quality data should be used as a signal amplifier: lowering the batch size converts cleaner data into more signal without amplifying noise. In the signal-limited regime,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25698","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/2605.25698/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":"2605.25698","created_at":"2026-05-26T02:04:51.107313+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.25698v1","created_at":"2026-05-26T02:04:51.107313+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25698","created_at":"2026-05-26T02:04:51.107313+00:00"},{"alias_kind":"pith_short_12","alias_value":"EB6HDFZLFDIA","created_at":"2026-05-26T02:04:51.107313+00:00"},{"alias_kind":"pith_short_16","alias_value":"EB6HDFZLFDIAXPE4","created_at":"2026-05-26T02:04:51.107313+00:00"},{"alias_kind":"pith_short_8","alias_value":"EB6HDFZL","created_at":"2026-05-26T02:04:51.107313+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/EB6HDFZLFDIAXPE4FZV6PTOP4Q","json":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q.json","graph_json":"https://pith.science/api/pith-number/EB6HDFZLFDIAXPE4FZV6PTOP4Q/graph.json","events_json":"https://pith.science/api/pith-number/EB6HDFZLFDIAXPE4FZV6PTOP4Q/events.json","paper":"https://pith.science/paper/EB6HDFZL"},"agent_actions":{"view_html":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q","download_json":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q.json","view_paper":"https://pith.science/paper/EB6HDFZL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.25698&json=true","fetch_graph":"https://pith.science/api/pith-number/EB6HDFZLFDIAXPE4FZV6PTOP4Q/graph.json","fetch_events":"https://pith.science/api/pith-number/EB6HDFZLFDIAXPE4FZV6PTOP4Q/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q/action/storage_attestation","attest_author":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q/action/author_attestation","sign_citation":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q/action/citation_signature","submit_replication":"https://pith.science/pith/EB6HDFZLFDIAXPE4FZV6PTOP4Q/action/replication_record"}},"created_at":"2026-05-26T02:04:51.107313+00:00","updated_at":"2026-05-26T02:04:51.107313+00:00"}