{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:PMIJFM6KJDOL6BKKAKDB4M6KCV","short_pith_number":"pith:PMIJFM6K","schema_version":"1.0","canonical_sha256":"7b1092b3ca48dcbf054a02861e33ca155ec497e10883f61ebe0a4353e2a80bab","source":{"kind":"arxiv","id":"2608.13277","version":1},"attestation_state":"computed","paper":{"title":"Mixture of Training: Recombining Small-Scale Scaffolded Pretraining Runs into a Larger Language Model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Keith Rush, Lucio Dery, Mohammed Sabry, Sean Augenstein","submitted_at":"2026-08-13T14:13:46Z","abstract_excerpt":"We ask whether language-model pre-training can be decomposed into smaller, independently trainable jobs that can later be recomposed into a coherent larger model. We introduce Mixture of Training (MoT), a scaffolded modular pre-training procedure that partitions a target Transformer into contiguous layer blocks, trains each block inside a frozen pretrained aligner scaffold, and then recomposes the trained blocks with an optional short end-to-end adaptation pass. On a 1.3B-parameter Gemma-style model trained on C4, MoT provides a small-scale proof of mechanism: independently trained depth slice"},"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":"2608.13277","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-13T14:13:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"5c31e59a5deda62866b7a86c18d2a354bcbc732457f005ee1c0daa87edde40c3","abstract_canon_sha256":"5928e0a5b0263697a4cf68c4c27e5281b19df8aaeef7d33cff1c055889b6cb17"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-14T01:00:50.258216Z","signature_b64":"nqhCvQSt5oRqJlfM3oVvz70cdyy14Se5FwwyT776v2dg52FqgfeMI7R3axepVpw3ugNC8Vf0rQ/+iRHA7mTlCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b1092b3ca48dcbf054a02861e33ca155ec497e10883f61ebe0a4353e2a80bab","last_reissued_at":"2026-08-14T01:00:50.256178Z","signature_status":"signed_v1","first_computed_at":"2026-08-14T01:00:50.256178Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mixture of Training: Recombining Small-Scale Scaffolded Pretraining Runs into a Larger Language Model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Keith Rush, Lucio Dery, Mohammed Sabry, Sean Augenstein","submitted_at":"2026-08-13T14:13:46Z","abstract_excerpt":"We ask whether language-model pre-training can be decomposed into smaller, independently trainable jobs that can later be recomposed into a coherent larger model. We introduce Mixture of Training (MoT), a scaffolded modular pre-training procedure that partitions a target Transformer into contiguous layer blocks, trains each block inside a frozen pretrained aligner scaffold, and then recomposes the trained blocks with an optional short end-to-end adaptation pass. On a 1.3B-parameter Gemma-style model trained on C4, MoT provides a small-scale proof of mechanism: independently trained depth slice"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.13277","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/2608.13277/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":"2608.13277","created_at":"2026-08-14T01:00:50.257206+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.13277v1","created_at":"2026-08-14T01:00:50.257206+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.13277","created_at":"2026-08-14T01:00:50.257206+00:00"},{"alias_kind":"pith_short_12","alias_value":"PMIJFM6KJDOL","created_at":"2026-08-14T01:00:50.257206+00:00"},{"alias_kind":"pith_short_16","alias_value":"PMIJFM6KJDOL6BKK","created_at":"2026-08-14T01:00:50.257206+00:00"},{"alias_kind":"pith_short_8","alias_value":"PMIJFM6K","created_at":"2026-08-14T01:00:50.257206+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/PMIJFM6KJDOL6BKKAKDB4M6KCV","json":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV.json","graph_json":"https://pith.science/api/pith-number/PMIJFM6KJDOL6BKKAKDB4M6KCV/graph.json","events_json":"https://pith.science/api/pith-number/PMIJFM6KJDOL6BKKAKDB4M6KCV/events.json","paper":"https://pith.science/paper/PMIJFM6K"},"agent_actions":{"view_html":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV","download_json":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV.json","view_paper":"https://pith.science/paper/PMIJFM6K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.13277&json=true","fetch_graph":"https://pith.science/api/pith-number/PMIJFM6KJDOL6BKKAKDB4M6KCV/graph.json","fetch_events":"https://pith.science/api/pith-number/PMIJFM6KJDOL6BKKAKDB4M6KCV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV/action/storage_attestation","attest_author":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV/action/author_attestation","sign_citation":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV/action/citation_signature","submit_replication":"https://pith.science/pith/PMIJFM6KJDOL6BKKAKDB4M6KCV/action/replication_record"}},"created_at":"2026-08-14T01:00:50.257206+00:00","updated_at":"2026-08-14T01:00:50.257206+00:00"}