{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "fedus-et-al-switch-transformers-2021",
    "title": "Fedus and colleagues: Switch Transformer sparse expert routing",
    "summary": "Simplified sparse mixture-of-experts routing for transformers with one expert selected per token and reported faster large-scale pre-training and trillion-parameter models.",
    "boundary": "This is the 2021 first preprint, later published in JMLR 2022; it does not invent MoE or imply that a trillion parameters are active per token.",
    "area": "Mixture-of-experts language models",
    "topic": "Artificial intelligence",
    "year": 2021,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "fedus-et-al-switch-transformers-2021",
    "target": {
      "record_id": "urn:uuid:50d312cd-f643-416e-a112-7473c98e905e",
      "version": 1,
      "content_hash": "sha256:d503132af14bdbbe32583e366e44be22fc7a46eb3dcee318c054be138ec327b3"
    },
    "humans": [
      "William Fedus",
      "Barret Zoph",
      "Noam Shazeer"
    ],
    "authors_complete": true,
    "human_role_note": "Complete named paper byline. Human research, source-reported system execution and experimental validation are distinct; individual effort shares are not measured.",
    "agents": [],
    "related": [],
    "subresults": [],
    "scientific_assessment": {
      "source_review": "SOURCE_CURATED",
      "scientific_review": "NOT_REVIEWED_BY_WRS",
      "formal_replay": "NOT_RUN",
      "provider_attestation": "NONE"
    },
    "sources": [
      {
        "id": "arxiv-switch-2101-03961",
        "title": "Fedus and colleagues: Switch Transformer sparse expert routing — original paper",
        "url": "https://arxiv.org/abs/2101.03961v3"
      },
      {
        "id": "google-switch-retrospective-2021",
        "title": "Fedus and colleagues: Switch Transformer sparse expert routing — company research source",
        "url": "https://research.google/blog/google-research-themes-from-2021-and-beyond/"
      }
    ],
    "points": "50000",
    "published_grade": {
      "id": "fedus-et-al-switch-transformers-2021",
      "entry_id": "fedus-et-al-switch-transformers-2021",
      "band": "A",
      "modifier": null,
      "target": {
        "record_id": "urn:uuid:50d312cd-f643-416e-a112-7473c98e905e",
        "version": 1,
        "content_hash": "sha256:d503132af14bdbbe32583e366e44be22fc7a46eb3dcee318c054be138ec327b3"
      },
      "what_changed": "Simplified sparse mixture-of-experts routing for transformers with one expert selected per token and reported faster large-scale pre-training and trillion-parameter models.",
      "scope": "Simplified sparse mixture-of-experts routing for transformers with one expert selected per token and reported faster large-scale pre-training and trillion-parameter models. Scope limitation: This is the 2021 first preprint, later published in JMLR 2022; it does not invent MoE or imply that a trillion parameters are active per token.",
      "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:50d312cd-f643-416e-a112-7473c98e905e",
        "version": 1,
        "content_hash": "sha256:d503132af14bdbbe32583e366e44be22fc7a46eb3dcee318c054be138ec327b3"
      },
      "edition": "ratings-e039106f30dec177"
    }
  },
  "version": {
    "kind": "PUBLIC_RECORD_VIEW",
    "previous": null,
    "published_at": "2026-09-27T00:00:00Z",
    "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": "William Fedus",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Fedus and colleagues: Switch Transformer sparse expert routing — original paper",
            "url": "https://arxiv.org/abs/2101.03961v3"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Barret Zoph",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Fedus and colleagues: Switch Transformer sparse expert routing — original paper",
            "url": "https://arxiv.org/abs/2101.03961v3"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Noam Shazeer",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Fedus and colleagues: Switch Transformer sparse expert routing — original paper",
            "url": "https://arxiv.org/abs/2101.03961v3"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Fedus and colleagues: Switch Transformer sparse expert routing — original paper",
            "url": "https://arxiv.org/abs/2101.03961v3"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-1",
          "kind": "source_reference",
          "source": {
            "citation": "Fedus and colleagues: Switch Transformer sparse expert routing — company research source",
            "url": "https://research.google/blog/google-research-themes-from-2021-and-beyond/"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "This is the 2021 first preprint, later published in JMLR 2022; it does not invent MoE or imply that a trillion parameters are active per token.",
        "domain": "Mixture-of-experts language models"
      },
      "sources": [
        {
          "citation": "Fedus and colleagues: Switch Transformer sparse expert routing — original paper",
          "url": "https://arxiv.org/abs/2101.03961v3"
        },
        {
          "citation": "Fedus and colleagues: Switch Transformer sparse expert routing — company research source",
          "url": "https://research.google/blog/google-research-themes-from-2021-and-beyond/"
        }
      ],
      "statement": "# Fedus and colleagues: Switch Transformer sparse expert routing\n\nSimplified sparse mixture-of-experts routing for transformers with one expert selected per token and reported faster large-scale pre-training and trillion-parameter models.\n\n## Scientific scope\n\nThis is the 2021 first preprint, later published in JMLR 2022; it does not invent MoE or imply that a trillion parameters are active per token.\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\nWilliam Fedus; Barret Zoph; Noam Shazeer\n\nComplete named paper byline. Human research, source-reported system execution and experimental validation are distinct; individual effort shares are not measured.\n\n\n\n## Original sources\n\n- [Fedus and colleagues: Switch Transformer sparse expert routing — original paper](https://arxiv.org/abs/2101.03961v3)\n- [Fedus and colleagues: Switch Transformer sparse expert routing — company research source](https://research.google/blog/google-research-themes-from-2021-and-beyond/)\n",
      "summary": "Simplified sparse mixture-of-experts routing for transformers with one expert selected per token and reported faster large-scale pre-training and trillion-parameter models.",
      "title": "Fedus and colleagues: Switch Transformer sparse expert routing"
    },
    "record_id": "urn:uuid:50d312cd-f643-416e-a112-7473c98e905e",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:50d312cd-f643-416e-a112-7473c98e905e",
      "version": 1,
      "content_hash": "sha256:d503132af14bdbbe32583e366e44be22fc7a46eb3dcee318c054be138ec327b3"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:a70868760d8438ceaf940cb31497ee3b961fc7cccef7e3ecc6cae3cef4386c90"
  },
  "target": {
    "record_id": "urn:uuid:50d312cd-f643-416e-a112-7473c98e905e",
    "version": 1,
    "content_hash": "sha256:d503132af14bdbbe32583e366e44be22fc7a46eb3dcee318c054be138ec327b3"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "fedus-et-al-switch-transformers-2021",
    "entry_id": "fedus-et-al-switch-transformers-2021",
    "revision": 1,
    "title": "Fedus and colleagues: Switch Transformer sparse expert routing",
    "context": "Mixture-of-experts language models",
    "track": "Human-designed AI research method or resource",
    "band": "A",
    "modifier": null,
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-09-27",
    "evidence_cutoff": "2026-09-27",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Simplified sparse mixture-of-experts routing for transformers with one expert selected per token and reported faster large-scale pre-training and trillion-parameter models.",
    "scope": "Simplified sparse mixture-of-experts routing for transformers with one expert selected per token and reported faster large-scale pre-training and trillion-parameter models. Scope limitation: This is the 2021 first preprint, later published in JMLR 2022; it does not invent MoE or imply that a trillion parameters are active per token.",
    "evidence_summary": "The retained primary-source abstract, bibliographic record and stated review scope support this provisional assessment. No independent experimental replication by WRS is claimed. This is the 2021 first preprint, later published in JMLR 2022; it does not invent MoE or imply that a trillion parameters are active per token.",
    "references": [
      {
        "title": "Fedus and colleagues: Switch Transformer sparse expert routing — original paper",
        "url": "https://arxiv.org/abs/2101.03961v3",
        "locator": "Abstract and complete bibliographic byline for 2101.03961v3; not an independent proof or experimental replication."
      },
      {
        "title": "Fedus and colleagues: Switch Transformer sparse expert routing — company research source",
        "url": "https://research.google/blog/google-research-themes-from-2021-and-beyond/",
        "locator": "Source-linked contribution and stated scientific limits; not an independent replication."
      }
    ],
    "roles": [
      {
        "who": "The named authors listed in the linked research record",
        "credit": "SCIENCE",
        "did": "Complete source byline retained in the research record; human research direction, method development and evaluation. Individual labor shares are not measured."
      }
    ],
    "target": {
      "record_id": "urn:uuid:50d312cd-f643-416e-a112-7473c98e905e",
      "version": 1,
      "content_hash": "sha256:d503132af14bdbbe32583e366e44be22fc7a46eb3dcee318c054be138ec327b3"
    },
    "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"
}