{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "yang-et-al-baichuan2-training-stability-resource-2023", "title": "Yang et al.: Baichuan2 training-stability and checkpoint resource", "summary": "Documented output-head normalization, maximum-logit regularization, elastic distributed-training choices and intermediate checkpoints for bilingual language-model training, enabling training-dynamics research.", "boundary": "Normalization and logit regularization are adaptations with cited predecessors, not invention of normalization, z-loss, ZeRO or positional encodings. Acquire general language-model training only; medical/legal benchmark sections do not establish clinical outcomes and are outside new intake. Model sizes share one package.", "area": "Language-model training stability and open checkpoints", "topic": "Artificial intelligence", "year": 2023, "lane": "HUMAN", "result": "EMPIRICAL", "package_id": "yang-et-al-baichuan2-training-stability-resource-2023", "target": { "record_id": "urn:uuid:d85e444e-a9a2-4482-a83e-ad023b2016e9", "version": 1, "content_hash": "sha256:fb9ca641ae0a5c6d7a75f447db47a76c7b6e63310ce48d9a7ffefd5be32b5172" }, "humans": [ "Aiyuan Yang", "Bin Xiao", "Bingning Wang", "Borong Zhang", "Chao Yin", "Chenxu Lv", "Da Pan", "Dian Wang", "Dong Yan", "Fan Yang", "Fei Deng", "Feng Wang", "Feng Liu", "Guangwei Ai", "Guosheng Dong", "Haizhou Zhao", "Hang Xu", "Haoze Sun", "Hongda Zhang", "Hui Liu", "Jiaming Ji", "Jian Xie", "Juntao Dai", "Kun Fang", "Lei Su", "Liang Song", "Lifeng Liu", "Liyun Ru", "Luyao Ma", "Mang Wang", "Mickel Liu", "MingAn Lin", "Nuolan Nie", "Peidong Guo", "Ruiyang Sun", "Tao Zhang", "Tianpeng Li", "Tianyu Li", "Wei Cheng", "Weipeng Chen", "Xiangrong Zeng", "Xiaochuan Wang", "Xiaoxi Chen", "Xin Men", "Xin Yu", "Xuehai Pan", "Yanjun Shen", "Yiding Wang", "Yiyu Li", "Youxin Jiang", "Yuchen Gao", "Yupeng Zhang", "Zenan Zhou", "Zhiying Wu" ], "authors_complete": true, "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.", "agents": [], "related": [], "subresults": [], "scientific_assessment": { "source_review": "SOURCE_CURATED", "scientific_review": "NOT_REVIEWED_BY_WRS", "formal_replay": "NOT_RUN", "provider_attestation": "NONE" }, "sources": [ { "id": "cur20261015-arxiv-2309-10305v1", "title": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" }, { "id": "cur20261015-support-yang-et-al-baichuan2-training-stability-resource-2023", "title": "Yang et al.: Baichuan2 training-stability and checkpoint resource — official research provenance", "url": "https://github.com/baichuan-inc/Baichuan2" } ], "points": "50000", "published_grade": { "id": "yang-et-al-baichuan2-training-stability-resource-2023", "entry_id": "yang-et-al-baichuan2-training-stability-resource-2023", "band": "B", "modifier": "+", "target": { "record_id": "urn:uuid:d85e444e-a9a2-4482-a83e-ad023b2016e9", "version": 1, "content_hash": "sha256:fb9ca641ae0a5c6d7a75f447db47a76c7b6e63310ce48d9a7ffefd5be32b5172" }, "what_changed": "Documented output-head normalization, maximum-logit regularization, elastic distributed-training choices and intermediate checkpoints for bilingual language-model training, enabling training-dynamics research.", "scope": "Documented output-head normalization, maximum-logit regularization, elastic distributed-training choices and intermediate checkpoints for bilingual language-model training, enabling training-dynamics research. Normalization and logit regularization are adaptations with cited predecessors, not invention of normalization, z-loss, ZeRO or positional encodings. Acquire general language-model training only; medical/legal benchmark sections do not establish clinical outcomes and are outside new intake. Model sizes share one package.", "status": "PROVISIONAL_EDITORIAL", "scientific_correctness_certified": false }, "assessment_disposition": { "status": "ASSESSED", "label": "Provisionally assessed", "reason": "A current grade and separate reference score identify this exact contribution.", "target": { "record_id": "urn:uuid:d85e444e-a9a2-4482-a83e-ad023b2016e9", "version": 1, "content_hash": "sha256:fb9ca641ae0a5c6d7a75f447db47a76c7b6e63310ce48d9a7ffefd5be32b5172" }, "edition": "ratings-e039106f30dec177" } }, "version": { "kind": "PUBLIC_RECORD_VIEW", "previous": null, "published_at": "2026-10-16T00:00:00Z", "publisher": { "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33", "role": "WRS_EDITORIAL_PUBLICATION" }, "record": { "contributors": [ { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Aiyuan Yang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Bin Xiao", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Bingning Wang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Borong Zhang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Chao Yin", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Chenxu Lv", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Da Pan", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Dian Wang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Dong Yan", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Fan Yang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Fei Deng", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Feng Wang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Feng Liu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Guangwei Ai", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Guosheng Dong", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Haizhou Zhao", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Hang Xu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Haoze Sun", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Hongda Zhang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Hui Liu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jiaming Ji", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jian Xie", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Juntao Dai", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Kun Fang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Lei Su", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Liang Song", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Lifeng Liu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Liyun Ru", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Luyao Ma", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Mang Wang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Mickel Liu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "discoverer" ], "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" } }, { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-1", "kind": "source_reference", "source": { "citation": "Yang et al.: Baichuan2 training-stability and checkpoint resource — official research provenance", "url": "https://github.com/baichuan-inc/Baichuan2" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "Normalization and logit regularization are adaptations with cited predecessors, not invention of normalization, z-loss, ZeRO or positional encodings. Acquire general language-model training only; medical/legal benchmark sections do not establish clinical outcomes and are outside new intake. Model sizes share one package.", "domain": "Language-model training stability and open checkpoints" }, "sources": [ { "citation": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1" }, { "citation": "Yang et al.: Baichuan2 training-stability and checkpoint resource — official research provenance", "url": "https://github.com/baichuan-inc/Baichuan2" } ], "statement": "# Yang et al.: Baichuan2 training-stability and checkpoint resource\n\nDocumented output-head normalization, maximum-logit regularization, elastic distributed-training choices and intermediate checkpoints for bilingual language-model training, enabling training-dynamics research.\n\n## Scientific scope\n\nNormalization and logit regularization are adaptations with cited predecessors, not invention of normalization, z-loss, ZeRO or positional encodings. Acquire general language-model training only; medical/legal benchmark sections do not establish clinical outcomes and are outside new intake. Model sizes share one package.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\nAiyuan Yang; Bin Xiao; Bingning Wang; Borong Zhang; Chao Yin; Chenxu Lv; Da Pan; Dian Wang; Dong Yan; Fan Yang; Fei Deng; Feng Wang; Feng Liu; Guangwei Ai; Guosheng Dong; Haizhou Zhao; Hang Xu; Haoze Sun; Hongda Zhang; Hui Liu; Jiaming Ji; Jian Xie; Juntao Dai; Kun Fang; Lei Su; Liang Song; Lifeng Liu; Liyun Ru; Luyao Ma; Mang Wang; Mickel Liu; MingAn Lin; Nuolan Nie; Peidong Guo; Ruiyang Sun; Tao Zhang; Tianpeng Li; Tianyu Li; Wei Cheng; Weipeng Chen; Xiangrong Zeng; Xiaochuan Wang; Xiaoxi Chen; Xin Men; Xin Yu; Xuehai Pan; Yanjun Shen; Yiding Wang; Yiyu Li; Youxin Jiang; Yuchen Gao; Yupeng Zhang; Zenan Zhou; Zhiying Wu\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [Baichuan 2: Open Large-scale Language Models](https://arxiv.org/abs/2309.10305v1)\n- [Yang et al.: Baichuan2 training-stability and checkpoint resource — official research provenance](https://github.com/baichuan-inc/Baichuan2)\n", "summary": "Documented output-head normalization, maximum-logit regularization, elastic distributed-training choices and intermediate checkpoints for bilingual language-model training, enabling training-dynamics research.", "title": "Yang et al.: Baichuan2 training-stability and checkpoint resource" }, "record_id": "urn:uuid:d85e444e-a9a2-4482-a83e-ad023b2016e9", "version": 1, "source_ref": { "record_id": "urn:uuid:d85e444e-a9a2-4482-a83e-ad023b2016e9", "version": 1, "content_hash": "sha256:fb9ca641ae0a5c6d7a75f447db47a76c7b6e63310ce48d9a7ffefd5be32b5172" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:f9249b703b6251b809461245a12f7c9645b2b397febd8b2c910eebbcd577460c" }, "target": { "record_id": "urn:uuid:d85e444e-a9a2-4482-a83e-ad023b2016e9", "version": 1, "content_hash": "sha256:fb9ca641ae0a5c6d7a75f447db47a76c7b6e63310ce48d9a7ffefd5be32b5172" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "yang-et-al-baichuan2-training-stability-resource-2023", "entry_id": "yang-et-al-baichuan2-training-stability-resource-2023", "revision": 1, "title": "Yang et al.: Baichuan2 training-stability and checkpoint resource", "context": "Language-model training stability and open checkpoints", "track": "Human-designed research method or resource", "band": "B", "modifier": "+", "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-16", "evidence_cutoff": "2026-10-16", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "Documented output-head normalization, maximum-logit regularization, elastic distributed-training choices and intermediate checkpoints for bilingual language-model training, enabling training-dynamics research.", "scope": "Documented output-head normalization, maximum-logit regularization, elastic distributed-training choices and intermediate checkpoints for bilingual language-model training, enabling training-dynamics research. Normalization and logit regularization are adaptations with cited predecessors, not invention of normalization, z-loss, ZeRO or positional encodings. Acquire general language-model training only; medical/legal benchmark sections do not establish clinical outcomes and are outside new intake. Model sizes share one package.", "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Normalization and logit regularization are adaptations with cited predecessors, not invention of normalization, z-loss, ZeRO or positional encodings. Acquire general language-model training only; medical/legal benchmark sections do not establish clinical outcomes and are outside new intake. Model sizes share one package.", "references": [ { "title": "Baichuan 2: Open Large-scale Language Models", "url": "https://arxiv.org/abs/2309.10305v1", "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification." }, { "title": "Yang et al.: Baichuan2 training-stability and checkpoint resource — official research provenance", "url": "https://github.com/baichuan-inc/Baichuan2", "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification." } ], "roles": [ { "who": "Source-listed authors and technical contributors", "credit": "SCIENCE", "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities." } ], "target": { "record_id": "urn:uuid:d85e444e-a9a2-4482-a83e-ad023b2016e9", "version": 1, "content_hash": "sha256:fb9ca641ae0a5c6d7a75f447db47a76c7b6e63310ce48d9a7ffefd5be32b5172" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "50000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": null, "reading_status": "SOURCE_LINKED_NOTE" }