{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:W56WP3TA26SXMTXX6KTGEVVPF4","short_pith_number":"pith:W56WP3TA","schema_version":"1.0","canonical_sha256":"b77d67ee60d7a5764ef7f2a66256af2f31c096441d1246da562d98eb1d25a825","source":{"kind":"arxiv","id":"2303.10464","version":2},"attestation_state":"computed","paper":{"title":"SPDF: Sparse Pre-training and Dense Fine-tuning for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Abhay Gupta, Dennis DeCoste, Kevin Leong, Sean Lie, Shreyas Saxena, Tianda Li, Vithursan Thangarasa, William Marshall","submitted_at":"2023-03-18T17:56:01Z","abstract_excerpt":"The pre-training and fine-tuning paradigm has contributed to a number of breakthroughs in Natural Language Processing (NLP). Instead of directly training on a downstream task, language models are first pre-trained on large datasets with cross-domain knowledge (e.g., Pile, MassiveText, etc.) and then fine-tuned on task-specific data (e.g., natural language generation, text summarization, etc.). Scaling the model and dataset size has helped improve the performance of LLMs, but unfortunately, this also lead to highly prohibitive computational costs. Pre-training LLMs often require orders of magni"},"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":"2303.10464","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-03-18T17:56:01Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"c94a7495c611c998d5b31d8dcb161b3ffd143eb692bab9710bada6c092679152","abstract_canon_sha256":"930ff175c866bd138759789d0ab18633fd41d48712b3b2d80628de3baa77fe8b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:35:49.550236Z","signature_b64":"lBke6iRbp6PiIZWV4tyHzcij+SCwP4or06z0TRECbBVfROhtHMjEkq1tmFQfHZI7E3zilGYpZ/rEqpaZalP5DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b77d67ee60d7a5764ef7f2a66256af2f31c096441d1246da562d98eb1d25a825","last_reissued_at":"2026-07-05T06:35:49.549818Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:35:49.549818Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SPDF: Sparse Pre-training and Dense Fine-tuning for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Abhay Gupta, Dennis DeCoste, Kevin Leong, Sean Lie, Shreyas Saxena, Tianda Li, Vithursan Thangarasa, William Marshall","submitted_at":"2023-03-18T17:56:01Z","abstract_excerpt":"The pre-training and fine-tuning paradigm has contributed to a number of breakthroughs in Natural Language Processing (NLP). Instead of directly training on a downstream task, language models are first pre-trained on large datasets with cross-domain knowledge (e.g., Pile, MassiveText, etc.) and then fine-tuned on task-specific data (e.g., natural language generation, text summarization, etc.). Scaling the model and dataset size has helped improve the performance of LLMs, but unfortunately, this also lead to highly prohibitive computational costs. Pre-training LLMs often require orders of magni"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.10464","kind":"arxiv","version":2},"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/2303.10464/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":"2303.10464","created_at":"2026-07-05T06:35:49.549874+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.10464v2","created_at":"2026-07-05T06:35:49.549874+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.10464","created_at":"2026-07-05T06:35:49.549874+00:00"},{"alias_kind":"pith_short_12","alias_value":"W56WP3TA26SX","created_at":"2026-07-05T06:35:49.549874+00:00"},{"alias_kind":"pith_short_16","alias_value":"W56WP3TA26SXMTXX","created_at":"2026-07-05T06:35:49.549874+00:00"},{"alias_kind":"pith_short_8","alias_value":"W56WP3TA","created_at":"2026-07-05T06:35:49.549874+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/W56WP3TA26SXMTXX6KTGEVVPF4","json":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4.json","graph_json":"https://pith.science/api/pith-number/W56WP3TA26SXMTXX6KTGEVVPF4/graph.json","events_json":"https://pith.science/api/pith-number/W56WP3TA26SXMTXX6KTGEVVPF4/events.json","paper":"https://pith.science/paper/W56WP3TA"},"agent_actions":{"view_html":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4","download_json":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4.json","view_paper":"https://pith.science/paper/W56WP3TA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.10464&json=true","fetch_graph":"https://pith.science/api/pith-number/W56WP3TA26SXMTXX6KTGEVVPF4/graph.json","fetch_events":"https://pith.science/api/pith-number/W56WP3TA26SXMTXX6KTGEVVPF4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4/action/storage_attestation","attest_author":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4/action/author_attestation","sign_citation":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4/action/citation_signature","submit_replication":"https://pith.science/pith/W56WP3TA26SXMTXX6KTGEVVPF4/action/replication_record"}},"created_at":"2026-07-05T06:35:49.549874+00:00","updated_at":"2026-07-05T06:35:49.549874+00:00"}