{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:GMKFNNXDP5RWKFLE6GFOGVXQBC","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"f0722d6de25663e7d09d87db38c7b273d1dfd257c9867c301967e7f4eb221f2e","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-21T21:27:48Z","title_canon_sha256":"81d174b60ba1bed421b79ed69dab96b2325952fb842b6e881a4991ed5157b435"},"schema_version":"1.0","source":{"id":"2502.15950","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.15950","created_at":"2026-07-05T10:18:18Z"},{"alias_kind":"arxiv_version","alias_value":"2502.15950v1","created_at":"2026-07-05T10:18:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.15950","created_at":"2026-07-05T10:18:18Z"},{"alias_kind":"pith_short_12","alias_value":"GMKFNNXDP5RW","created_at":"2026-07-05T10:18:18Z"},{"alias_kind":"pith_short_16","alias_value":"GMKFNNXDP5RWKFLE","created_at":"2026-07-05T10:18:18Z"},{"alias_kind":"pith_short_8","alias_value":"GMKFNNXD","created_at":"2026-07-05T10:18:18Z"}],"graph_snapshots":[{"event_id":"sha256:dce2fa2a666f40d460c9b62a3d9c3e2199af32e6536fcdb58527388e408adc61","target":"graph","created_at":"2026-07-05T10:18:18Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2502.15950/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We propose a method to optimize language model pre-training data mixtures through efficient approximation of the cross-entropy loss corresponding to each candidate mixture via a Mixture of Data Experts (MDE). We use this approximation as a source of additional features in a regression model, trained from observations of model loss for a small number of mixtures.\n  Experiments with Transformer decoder-only language models in the range of 70M to 1B parameters on the SlimPajama dataset show that our method achieves significantly better performance than approaches that train regression models usin","authors_text":"Alekh Agarwal, Kristina Toutanova, Lior Belenki, Tianze Shi","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-21T21:27:48Z","title":"Optimizing Pre-Training Data Mixtures with Mixtures of Data Expert Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.15950","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:99a2ade6258b6e9f11649a4f6acdd76cee2b56431324fb984b99035b5ecbc6cb","target":"record","created_at":"2026-07-05T10:18:18Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"f0722d6de25663e7d09d87db38c7b273d1dfd257c9867c301967e7f4eb221f2e","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-21T21:27:48Z","title_canon_sha256":"81d174b60ba1bed421b79ed69dab96b2325952fb842b6e881a4991ed5157b435"},"schema_version":"1.0","source":{"id":"2502.15950","kind":"arxiv","version":1}},"canonical_sha256":"331456b6e37f63651564f18ae356f0088781b6f0517184238f6b841166f3054d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"331456b6e37f63651564f18ae356f0088781b6f0517184238f6b841166f3054d","first_computed_at":"2026-07-05T10:18:18.907222Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:18:18.907222Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"r/pCzeZD/u/kAm8d7Vytatl/MgsDq/XUd9Wvpfc7Fyn0xI1uAweJxAkuhQeVPSDB47L44S8zSg91GQE2WyIzAw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:18:18.907665Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.15950","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:99a2ade6258b6e9f11649a4f6acdd76cee2b56431324fb984b99035b5ecbc6cb","sha256:dce2fa2a666f40d460c9b62a3d9c3e2199af32e6536fcdb58527388e408adc61"],"state_sha256":"095bba51b6d96fc90d76246bc5ab43caaf2423514f911addbb1ff4bef8470a47"}