{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:N7GCV265G4PZNBQEB654QR636O","short_pith_number":"pith:N7GCV265","schema_version":"1.0","canonical_sha256":"6fcc2aebdd371f9686040fbbc847dbf38d9ee478c105331c0bae01fd2c7a5d3e","source":{"kind":"arxiv","id":"2410.16208","version":4},"attestation_state":"computed","paper":{"title":"Compute-Constrained Data Selection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Alexander M. Rush, Junjie Oscar Yin","submitted_at":"2024-10-21T17:11:21Z","abstract_excerpt":"Data selection can reduce the amount of training data needed to finetune LLMs; however, the efficacy of data selection scales directly with its compute. Motivated by the practical challenge of compute-constrained finetuning, we consider the setting in which both the cost of selecting data and training are budgeted for. We first formalize the problem of data selection with a cost-aware utility function, and model the data selection problem as trading off initial-selection cost for training gain. We run a comprehensive sweep of experiments across multiple tasks, varying compute budget by scaling"},"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":"2410.16208","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-21T17:11:21Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"63602bde42e74168df4e3bc6f9399585014689c66a5f3920e60765551961ac1c","abstract_canon_sha256":"1c42ad511ec1bb418032054690619acdd4b9915f7a11795be397c4cf2d449e72"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:45:38.345435Z","signature_b64":"kCqVbCl8uVA2TPFFbSTjmrFP5612hDaxvYKpkAr+rddLBAmMTikwdaAawc7I1VrcD0XmhFP+Ai0QnXKXRiK7Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6fcc2aebdd371f9686040fbbc847dbf38d9ee478c105331c0bae01fd2c7a5d3e","last_reissued_at":"2026-07-05T10:45:38.344899Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:45:38.344899Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Compute-Constrained Data Selection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Alexander M. Rush, Junjie Oscar Yin","submitted_at":"2024-10-21T17:11:21Z","abstract_excerpt":"Data selection can reduce the amount of training data needed to finetune LLMs; however, the efficacy of data selection scales directly with its compute. Motivated by the practical challenge of compute-constrained finetuning, we consider the setting in which both the cost of selecting data and training are budgeted for. We first formalize the problem of data selection with a cost-aware utility function, and model the data selection problem as trading off initial-selection cost for training gain. We run a comprehensive sweep of experiments across multiple tasks, varying compute budget by scaling"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.16208","kind":"arxiv","version":4},"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/2410.16208/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":"2410.16208","created_at":"2026-07-05T10:45:38.344966+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.16208v4","created_at":"2026-07-05T10:45:38.344966+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.16208","created_at":"2026-07-05T10:45:38.344966+00:00"},{"alias_kind":"pith_short_12","alias_value":"N7GCV265G4PZ","created_at":"2026-07-05T10:45:38.344966+00:00"},{"alias_kind":"pith_short_16","alias_value":"N7GCV265G4PZNBQE","created_at":"2026-07-05T10:45:38.344966+00:00"},{"alias_kind":"pith_short_8","alias_value":"N7GCV265","created_at":"2026-07-05T10:45:38.344966+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.11810","citing_title":"GRACE: A Dynamic Coreset Selection Framework for Large Language Model Optimization","ref_index":78,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05227","citing_title":"Rethinking Data Curation in LLM Training: Online Reweighting Offers Better Generalization than Offline Methods","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O","json":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O.json","graph_json":"https://pith.science/api/pith-number/N7GCV265G4PZNBQEB654QR636O/graph.json","events_json":"https://pith.science/api/pith-number/N7GCV265G4PZNBQEB654QR636O/events.json","paper":"https://pith.science/paper/N7GCV265"},"agent_actions":{"view_html":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O","download_json":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O.json","view_paper":"https://pith.science/paper/N7GCV265","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.16208&json=true","fetch_graph":"https://pith.science/api/pith-number/N7GCV265G4PZNBQEB654QR636O/graph.json","fetch_events":"https://pith.science/api/pith-number/N7GCV265G4PZNBQEB654QR636O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O/action/storage_attestation","attest_author":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O/action/author_attestation","sign_citation":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O/action/citation_signature","submit_replication":"https://pith.science/pith/N7GCV265G4PZNBQEB654QR636O/action/replication_record"}},"created_at":"2026-07-05T10:45:38.344966+00:00","updated_at":"2026-07-05T10:45:38.344966+00:00"}