{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:X3YCNXYZ6AY2MCTTUEHWP5X6HS","short_pith_number":"pith:X3YCNXYZ","schema_version":"1.0","canonical_sha256":"bef026df19f031a60a73a10f67f6fe3cabef9e596d93e773c7c6bd255b22b3fa","source":{"kind":"arxiv","id":"2606.29554","version":1},"attestation_state":"computed","paper":{"title":"Optimizer Memory Makes Shuffle Order a First-Order Source of Fine-Tuning Noise","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NA","math.NA","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"John Sweeney","submitted_at":"2026-06-28T18:44:08Z","abstract_excerpt":"Shuffle order can be a larger source of fine-tuning noise than a memoryless analysis predicts: fixed-clock optimizer memory makes local equal-multiset contrasts first order in the learning rate rather than second order, and the resulting order channel can be large enough for a single seed to flip a close A/B comparison. We isolate this mechanism and derive a fit-free way to size the noise it produces. For a memoryless optimizer, reordering an equal multiset has no first-order endpoint term; the leading local contrast is the $O(\\eta^2)$ gradient bracket. Fixed-clock optimizers such as AdamW are"},"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":"2606.29554","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-28T18:44:08Z","cross_cats_sorted":["cs.NA","math.NA","math.OC","stat.ML"],"title_canon_sha256":"1df97dd85af081967e10fa5e5ed5f7aa6660848e54e8e393a55eb9399fbb1aec","abstract_canon_sha256":"1aac164232858b791529ac027e5219f477f1da3c2c782b7964f1df6c55d4db35"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T01:18:11.764194Z","signature_b64":"rJCxI3q57lZ8JM/CXuaA5g/jrsVBczTAH5HFTKyvzu760rmPrQ0h0uxFUXgaucIjqCnuYnm8X4LPU7q0FLwoAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bef026df19f031a60a73a10f67f6fe3cabef9e596d93e773c7c6bd255b22b3fa","last_reissued_at":"2026-06-30T01:18:11.763628Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T01:18:11.763628Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimizer Memory Makes Shuffle Order a First-Order Source of Fine-Tuning Noise","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NA","math.NA","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"John Sweeney","submitted_at":"2026-06-28T18:44:08Z","abstract_excerpt":"Shuffle order can be a larger source of fine-tuning noise than a memoryless analysis predicts: fixed-clock optimizer memory makes local equal-multiset contrasts first order in the learning rate rather than second order, and the resulting order channel can be large enough for a single seed to flip a close A/B comparison. We isolate this mechanism and derive a fit-free way to size the noise it produces. For a memoryless optimizer, reordering an equal multiset has no first-order endpoint term; the leading local contrast is the $O(\\eta^2)$ gradient bracket. Fixed-clock optimizers such as AdamW are"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.29554","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/2606.29554/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":"2606.29554","created_at":"2026-06-30T01:18:11.763710+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.29554v1","created_at":"2026-06-30T01:18:11.763710+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.29554","created_at":"2026-06-30T01:18:11.763710+00:00"},{"alias_kind":"pith_short_12","alias_value":"X3YCNXYZ6AY2","created_at":"2026-06-30T01:18:11.763710+00:00"},{"alias_kind":"pith_short_16","alias_value":"X3YCNXYZ6AY2MCTT","created_at":"2026-06-30T01:18:11.763710+00:00"},{"alias_kind":"pith_short_8","alias_value":"X3YCNXYZ","created_at":"2026-06-30T01:18:11.763710+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/X3YCNXYZ6AY2MCTTUEHWP5X6HS","json":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS.json","graph_json":"https://pith.science/api/pith-number/X3YCNXYZ6AY2MCTTUEHWP5X6HS/graph.json","events_json":"https://pith.science/api/pith-number/X3YCNXYZ6AY2MCTTUEHWP5X6HS/events.json","paper":"https://pith.science/paper/X3YCNXYZ"},"agent_actions":{"view_html":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS","download_json":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS.json","view_paper":"https://pith.science/paper/X3YCNXYZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.29554&json=true","fetch_graph":"https://pith.science/api/pith-number/X3YCNXYZ6AY2MCTTUEHWP5X6HS/graph.json","fetch_events":"https://pith.science/api/pith-number/X3YCNXYZ6AY2MCTTUEHWP5X6HS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS/action/storage_attestation","attest_author":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS/action/author_attestation","sign_citation":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS/action/citation_signature","submit_replication":"https://pith.science/pith/X3YCNXYZ6AY2MCTTUEHWP5X6HS/action/replication_record"}},"created_at":"2026-06-30T01:18:11.763710+00:00","updated_at":"2026-06-30T01:18:11.763710+00:00"}