{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:OOWOVAOJWULVAHDCY3NUMDVWRC","short_pith_number":"pith:OOWOVAOJ","schema_version":"1.0","canonical_sha256":"73acea81c9b517501c62c6db460eb68890f2d7c0fb81294009e1de97ea9b810c","source":{"kind":"arxiv","id":"2408.16947","version":1},"attestation_state":"computed","paper":{"title":"An Empirical Study of Scaling Laws for Transfer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Matthew Barnett","submitted_at":"2024-08-30T00:06:29Z","abstract_excerpt":"We present a limited empirical study of scaling laws for transfer learning in transformer models. More specifically, we examine a scaling law that incorporates a \"transfer gap\" term, indicating the effectiveness of pre-training on one distribution when optimizing for downstream performance on another distribution. When the transfer gap is low, pre-training is a cost-effective strategy for improving downstream performance. Conversely, when the gap is high, collecting high-quality fine-tuning data becomes relatively more cost effective. Fitting the scaling law to experiments from diverse dataset"},"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":"2408.16947","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-30T00:06:29Z","cross_cats_sorted":[],"title_canon_sha256":"1fb1e87a64f6c9c5123a5a8d1fa60cc2d8aa5382e5a53441b450354653cc5a2d","abstract_canon_sha256":"9dd4336774ae08d7c1a24ba4d7226099b34eba2f582f8974951c44fbee28299f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:01:08.662923Z","signature_b64":"+LL0HgdTgEMvHLPdZqF/vZiKgguSAj93RJMFGADozxuf8mPpj9x/c4SMkqkDyAAz4rrpPGaeJdfzbovLoBvgCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"73acea81c9b517501c62c6db460eb68890f2d7c0fb81294009e1de97ea9b810c","last_reissued_at":"2026-07-05T09:01:08.662538Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:01:08.662538Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Empirical Study of Scaling Laws for Transfer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Matthew Barnett","submitted_at":"2024-08-30T00:06:29Z","abstract_excerpt":"We present a limited empirical study of scaling laws for transfer learning in transformer models. More specifically, we examine a scaling law that incorporates a \"transfer gap\" term, indicating the effectiveness of pre-training on one distribution when optimizing for downstream performance on another distribution. When the transfer gap is low, pre-training is a cost-effective strategy for improving downstream performance. Conversely, when the gap is high, collecting high-quality fine-tuning data becomes relatively more cost effective. Fitting the scaling law to experiments from diverse dataset"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.16947","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/2408.16947/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":"2408.16947","created_at":"2026-07-05T09:01:08.662606+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.16947v1","created_at":"2026-07-05T09:01:08.662606+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.16947","created_at":"2026-07-05T09:01:08.662606+00:00"},{"alias_kind":"pith_short_12","alias_value":"OOWOVAOJWULV","created_at":"2026-07-05T09:01:08.662606+00:00"},{"alias_kind":"pith_short_16","alias_value":"OOWOVAOJWULVAHDC","created_at":"2026-07-05T09:01:08.662606+00:00"},{"alias_kind":"pith_short_8","alias_value":"OOWOVAOJ","created_at":"2026-07-05T09:01:08.662606+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.10378","citing_title":"Discovering Hierarchical Latent Capabilities of Language Models via Causal Representation Learning","ref_index":10,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC","json":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC.json","graph_json":"https://pith.science/api/pith-number/OOWOVAOJWULVAHDCY3NUMDVWRC/graph.json","events_json":"https://pith.science/api/pith-number/OOWOVAOJWULVAHDCY3NUMDVWRC/events.json","paper":"https://pith.science/paper/OOWOVAOJ"},"agent_actions":{"view_html":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC","download_json":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC.json","view_paper":"https://pith.science/paper/OOWOVAOJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.16947&json=true","fetch_graph":"https://pith.science/api/pith-number/OOWOVAOJWULVAHDCY3NUMDVWRC/graph.json","fetch_events":"https://pith.science/api/pith-number/OOWOVAOJWULVAHDCY3NUMDVWRC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC/action/storage_attestation","attest_author":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC/action/author_attestation","sign_citation":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC/action/citation_signature","submit_replication":"https://pith.science/pith/OOWOVAOJWULVAHDCY3NUMDVWRC/action/replication_record"}},"created_at":"2026-07-05T09:01:08.662606+00:00","updated_at":"2026-07-05T09:01:08.662606+00:00"}