{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:XNW4OZZIBC3TQMR5GBAKPMN2AD","short_pith_number":"pith:XNW4OZZI","schema_version":"1.0","canonical_sha256":"bb6dc7672808b738323d3040a7b1ba00eb0169e1d771fb1bdd9592eb98d2b368","source":{"kind":"arxiv","id":"2310.15389","version":1},"attestation_state":"computed","paper":{"title":"Irreducible Curriculum for Language Model Pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Martin Jaggi, Simin Fan","submitted_at":"2023-10-23T22:41:33Z","abstract_excerpt":"Automatic data selection and curriculum design for training large language models is challenging, with only a few existing methods showing improvements over standard training. Furthermore, current schemes focus on domain-level selection, overlooking the more fine-grained contributions of each individual training point. It is difficult to apply traditional datapoint selection methods on large language models: most online batch selection methods perform two-times forward or backward passes, which introduces considerable extra costs with large-scale models. To mitigate these obstacles, we propose"},"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":"2310.15389","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-23T22:41:33Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"0d128fc0b94857848d71fd4e2e7a323c3c750b1b5a9a246fbbf700ab11af946f","abstract_canon_sha256":"43147d6da2b90b4f5e1376846befb4c4db0a6ca9a862c73e5d1bc988b245f703"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:04:25.411515Z","signature_b64":"WtTME/yEznTdLLuKqI/RdNcj7Ve9EFwENi6EbWaxK79LWL/+HYdhxyHqw+zF5YJzBxGgSo2iwW0CmqgH9anKCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bb6dc7672808b738323d3040a7b1ba00eb0169e1d771fb1bdd9592eb98d2b368","last_reissued_at":"2026-07-05T07:04:25.410955Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:04:25.410955Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Irreducible Curriculum for Language Model Pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Martin Jaggi, Simin Fan","submitted_at":"2023-10-23T22:41:33Z","abstract_excerpt":"Automatic data selection and curriculum design for training large language models is challenging, with only a few existing methods showing improvements over standard training. Furthermore, current schemes focus on domain-level selection, overlooking the more fine-grained contributions of each individual training point. It is difficult to apply traditional datapoint selection methods on large language models: most online batch selection methods perform two-times forward or backward passes, which introduces considerable extra costs with large-scale models. To mitigate these obstacles, we propose"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.15389","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/2310.15389/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":"2310.15389","created_at":"2026-07-05T07:04:25.411014+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.15389v1","created_at":"2026-07-05T07:04:25.411014+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.15389","created_at":"2026-07-05T07:04:25.411014+00:00"},{"alias_kind":"pith_short_12","alias_value":"XNW4OZZIBC3T","created_at":"2026-07-05T07:04:25.411014+00:00"},{"alias_kind":"pith_short_16","alias_value":"XNW4OZZIBC3TQMR5","created_at":"2026-07-05T07:04:25.411014+00:00"},{"alias_kind":"pith_short_8","alias_value":"XNW4OZZI","created_at":"2026-07-05T07:04:25.411014+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2502.10248","citing_title":"Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model","ref_index":262,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26687","citing_title":"COPUS: Co-adaptive Parallelism and Batch Size Selection in Large Language Model Training","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD","json":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD.json","graph_json":"https://pith.science/api/pith-number/XNW4OZZIBC3TQMR5GBAKPMN2AD/graph.json","events_json":"https://pith.science/api/pith-number/XNW4OZZIBC3TQMR5GBAKPMN2AD/events.json","paper":"https://pith.science/paper/XNW4OZZI"},"agent_actions":{"view_html":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD","download_json":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD.json","view_paper":"https://pith.science/paper/XNW4OZZI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.15389&json=true","fetch_graph":"https://pith.science/api/pith-number/XNW4OZZIBC3TQMR5GBAKPMN2AD/graph.json","fetch_events":"https://pith.science/api/pith-number/XNW4OZZIBC3TQMR5GBAKPMN2AD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD/action/storage_attestation","attest_author":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD/action/author_attestation","sign_citation":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD/action/citation_signature","submit_replication":"https://pith.science/pith/XNW4OZZIBC3TQMR5GBAKPMN2AD/action/replication_record"}},"created_at":"2026-07-05T07:04:25.411014+00:00","updated_at":"2026-07-05T07:04:25.411014+00:00"}