{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2XLH6DKMQDHJKJN7YLDBKEDFLH","short_pith_number":"pith:2XLH6DKM","schema_version":"1.0","canonical_sha256":"d5d67f0d4c80ce9525bfc2c615106559cbe364b4688f8588d72e7407339a5fdb","source":{"kind":"arxiv","id":"2406.04391","version":2},"attestation_state":"computed","paper":{"title":"Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Adam Ibrahim, Brando Miranda, Gabriel Mukobi, Hailey Schoelkopf, Herbie Bradley, Rylan Schaeffer, Sanmi Koyejo, Stella Biderman, Varun Madan","submitted_at":"2024-06-06T17:46:56Z","abstract_excerpt":"Predicting changes from scaling advanced AI systems is a desirable property for engineers, economists, governments and industry alike, and, while a well-established literature exists on how pretraining performance scales, predictable scaling behavior on downstream capabilities remains elusive. While many factors are certainly responsible, this paper identifies a significant factor that makes predicting scaling behavior on widely used multiple-choice question answering benchmarks challenging and illuminates a path towards making such downstream evaluations predictable with scale. Using five mod"},"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":"2406.04391","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-06T17:46:56Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"a35c89bee64e0a5b65f073928db1cedcff3460ef1d73be9127623a2f4a50cb2c","abstract_canon_sha256":"78cc358d4dbecf93fee907d2cfb4844377eb76669b2c06a59cda3c097fab79ef"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:10:23.806637Z","signature_b64":"XCIicZCVDUD4WsApqZ83a3TkjBRtJEpYko+jt00csObj6d6ZoAAtr3dArS6CoS5DwqSK7gBuROftWaB75nm8Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d5d67f0d4c80ce9525bfc2c615106559cbe364b4688f8588d72e7407339a5fdb","last_reissued_at":"2026-07-05T10:10:23.806146Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:10:23.806146Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Adam Ibrahim, Brando Miranda, Gabriel Mukobi, Hailey Schoelkopf, Herbie Bradley, Rylan Schaeffer, Sanmi Koyejo, Stella Biderman, Varun Madan","submitted_at":"2024-06-06T17:46:56Z","abstract_excerpt":"Predicting changes from scaling advanced AI systems is a desirable property for engineers, economists, governments and industry alike, and, while a well-established literature exists on how pretraining performance scales, predictable scaling behavior on downstream capabilities remains elusive. While many factors are certainly responsible, this paper identifies a significant factor that makes predicting scaling behavior on widely used multiple-choice question answering benchmarks challenging and illuminates a path towards making such downstream evaluations predictable with scale. Using five mod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.04391","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/2406.04391/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":"2406.04391","created_at":"2026-07-05T10:10:23.806206+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.04391v2","created_at":"2026-07-05T10:10:23.806206+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.04391","created_at":"2026-07-05T10:10:23.806206+00:00"},{"alias_kind":"pith_short_12","alias_value":"2XLH6DKMQDHJ","created_at":"2026-07-05T10:10:23.806206+00:00"},{"alias_kind":"pith_short_16","alias_value":"2XLH6DKMQDHJKJN7","created_at":"2026-07-05T10:10:23.806206+00:00"},{"alias_kind":"pith_short_8","alias_value":"2XLH6DKM","created_at":"2026-07-05T10:10:23.806206+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.02464","citing_title":"Will Scaling Improve Social Simulation with LLMs?","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07616","citing_title":"Item Response Scaling Laws: A Measurement Theory Approach for Efficient and Generalizable Neural Scaling Estimation","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2502.12120","citing_title":"LLMs on the Line: Data Determines Loss-to-Loss Scaling Laws","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH","json":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH.json","graph_json":"https://pith.science/api/pith-number/2XLH6DKMQDHJKJN7YLDBKEDFLH/graph.json","events_json":"https://pith.science/api/pith-number/2XLH6DKMQDHJKJN7YLDBKEDFLH/events.json","paper":"https://pith.science/paper/2XLH6DKM"},"agent_actions":{"view_html":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH","download_json":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH.json","view_paper":"https://pith.science/paper/2XLH6DKM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.04391&json=true","fetch_graph":"https://pith.science/api/pith-number/2XLH6DKMQDHJKJN7YLDBKEDFLH/graph.json","fetch_events":"https://pith.science/api/pith-number/2XLH6DKMQDHJKJN7YLDBKEDFLH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH/action/storage_attestation","attest_author":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH/action/author_attestation","sign_citation":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH/action/citation_signature","submit_replication":"https://pith.science/pith/2XLH6DKMQDHJKJN7YLDBKEDFLH/action/replication_record"}},"created_at":"2026-07-05T10:10:23.806206+00:00","updated_at":"2026-07-05T10:10:23.806206+00:00"}