{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QMCAWIL5AXFP4Y6GTLMEHQUJT2","short_pith_number":"pith:QMCAWIL5","schema_version":"1.0","canonical_sha256":"83040b217d05cafe63c69ad843c2899e9fff1b2b8c52e61e1b0ba5ea911e6b9a","source":{"kind":"arxiv","id":"2604.09258","version":2},"attestation_state":"computed","paper":{"title":"Nexus: Same Pretraining Loss, Better Downstream Generalization via Common Minima","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Converging to common minima across data sources during pretraining improves downstream generalization even at identical loss values.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Huanran Chen, Huaqing Zhang, Jun Zhu, Ke Shen, Xiao Li, Yinpeng Dong","submitted_at":"2026-04-10T12:17:18Z","abstract_excerpt":"The foundational capabilities of large language models are acquired during pretraining on internet-scale, highly heterogeneous data mixtures. In this work, we investigate an interesting geometric question regarding the converged state of pretraining: Does the model converge to a common minimizer across all data sources (e.g., \\cref{fig:cwa_illustration:close}), or merely a minimizer of the summed loss (e.g., \\cref{fig:cwa_illustration:distant})? We hypothesize that the geometric \"closeness\" of task-specific minima is intrinsically linked to downstream generalization. We reveal that standard op"},"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":true},"canonical_record":{"source":{"id":"2604.09258","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-10T12:17:18Z","cross_cats_sorted":[],"title_canon_sha256":"f59b06e55ceb072b5bc53af46f96114d78f57fc9ea1d72a33a834856408e498c","abstract_canon_sha256":"84a5948012050bc97e76c01cda4ff86e78a223497cc5590b26437950090b5b33"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:39.855575Z","signature_b64":"I0RiaycFY+HivsYTP/WyrRRXj6I0mkCpUSXQDSjvcOFrtW2BNmPqM9cPPMdGFz/1pMcKHwRVYN20gPIYAKhFBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"83040b217d05cafe63c69ad843c2899e9fff1b2b8c52e61e1b0ba5ea911e6b9a","last_reissued_at":"2026-05-28T01:04:39.854864Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:39.854864Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Nexus: Same Pretraining Loss, Better Downstream Generalization via Common Minima","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Converging to common minima across data sources during pretraining improves downstream generalization even at identical loss values.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Huanran Chen, Huaqing Zhang, Jun Zhu, Ke Shen, Xiao Li, Yinpeng Dong","submitted_at":"2026-04-10T12:17:18Z","abstract_excerpt":"The foundational capabilities of large language models are acquired during pretraining on internet-scale, highly heterogeneous data mixtures. In this work, we investigate an interesting geometric question regarding the converged state of pretraining: Does the model converge to a common minimizer across all data sources (e.g., \\cref{fig:cwa_illustration:close}), or merely a minimizer of the summed loss (e.g., \\cref{fig:cwa_illustration:distant})? We hypothesize that the geometric \"closeness\" of task-specific minima is intrinsically linked to downstream generalization. We reveal that standard op"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Nexus significantly boosts downstream performance, despite achieving the same pretraining loss. Notably, on the 3B model, Nexus reduces the out-of-distribution loss by 0.012 and yields up to a 15.0% accuracy improvement on complex reasoning tasks (e.g., GSM8k).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The geometric closeness of task-specific minima is intrinsically linked to downstream generalization, and that maximizing gradient similarity during optimization produces this closeness.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Nexus optimizer improves LLM downstream performance by converging to common minima across data sources despite identical pretraining loss.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Converging to common minima across data sources during pretraining improves downstream generalization even at identical loss values.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"a584887b95800e03e044bfd266dd9f91be2c62b2cc57f44f5fd665782b17d358"},"source":{"id":"2604.09258","kind":"arxiv","version":2},"verdict":{"id":"d3ba602d-47e6-4b65-9d4a-d0b8d58839f6","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T17:50:26.241619Z","strongest_claim":"Nexus significantly boosts downstream performance, despite achieving the same pretraining loss. Notably, on the 3B model, Nexus reduces the out-of-distribution loss by 0.012 and yields up to a 15.0% accuracy improvement on complex reasoning tasks (e.g., GSM8k).","one_line_summary":"Nexus optimizer improves LLM downstream performance by converging to common minima across data sources despite identical pretraining loss.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The geometric closeness of task-specific minima is intrinsically linked to downstream generalization, and that maximizing gradient similarity during optimization produces this closeness.","pith_extraction_headline":"Converging to common minima across data sources during pretraining improves downstream generalization even at identical loss values."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.09258/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":2,"snapshot_sha256":"061188b7996e67bdbad99c3626f52832ce173c1e1b7cc37dd2c29826c484c4aa"},"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":"2604.09258","created_at":"2026-05-28T01:04:39.854936+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.09258v2","created_at":"2026-05-28T01:04:39.854936+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.09258","created_at":"2026-05-28T01:04:39.854936+00:00"},{"alias_kind":"pith_short_12","alias_value":"QMCAWIL5AXFP","created_at":"2026-05-28T01:04:39.854936+00:00"},{"alias_kind":"pith_short_16","alias_value":"QMCAWIL5AXFP4Y6G","created_at":"2026-05-28T01:04:39.854936+00:00"},{"alias_kind":"pith_short_8","alias_value":"QMCAWIL5","created_at":"2026-05-28T01:04:39.854936+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.20134","citing_title":"TrajTok: Adaptive Spatial Tokenization for Trajectory Representation Learning","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2","json":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2.json","graph_json":"https://pith.science/api/pith-number/QMCAWIL5AXFP4Y6GTLMEHQUJT2/graph.json","events_json":"https://pith.science/api/pith-number/QMCAWIL5AXFP4Y6GTLMEHQUJT2/events.json","paper":"https://pith.science/paper/QMCAWIL5"},"agent_actions":{"view_html":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2","download_json":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2.json","view_paper":"https://pith.science/paper/QMCAWIL5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.09258&json=true","fetch_graph":"https://pith.science/api/pith-number/QMCAWIL5AXFP4Y6GTLMEHQUJT2/graph.json","fetch_events":"https://pith.science/api/pith-number/QMCAWIL5AXFP4Y6GTLMEHQUJT2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2/action/storage_attestation","attest_author":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2/action/author_attestation","sign_citation":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2/action/citation_signature","submit_replication":"https://pith.science/pith/QMCAWIL5AXFP4Y6GTLMEHQUJT2/action/replication_record"}},"created_at":"2026-05-28T01:04:39.854936+00:00","updated_at":"2026-05-28T01:04:39.854936+00:00"}