{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "lee-et-al-gecko-2024",
    "title": "Gecko: task/query synthesis and passage relabeling for embedding distillation",
    "summary": "Generated task descriptions and queries from web passages, retrieved alternative passages and used an LLM to relabel positives/hard negatives before contrastive embedding fine-tuning.",
    "boundary": "The seed passage need not remain the best positive. Synthetic labels and few-shot relevance/query-likelihood judgments are not ground-truth guarantees. Teacher/data-generation and pretrained-backbone costs remain; dimension and leaderboard comparisons are source snapshots, not universal compression or current-rank claims.",
    "area": "Text embedding training",
    "topic": "Artificial intelligence",
    "year": 2024,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "lee-et-al-gecko-2024",
    "target": {
      "record_id": "urn:uuid:35aee8ee-595d-4579-acac-ad0cc3a004d9",
      "version": 1,
      "content_hash": "sha256:35f4955a1d68192e17410de259ef04825fbfc9a60102a80f34a580d731457f10"
    },
    "humans": [
      "Jinhyuk Lee",
      "Zhuyun Dai",
      "Xiaoqi Ren",
      "Blair Chen",
      "Daniel Cer",
      "Jeremy R. Cole",
      "Kai Hui",
      "Michael Boratko",
      "Rajvi Kapadia",
      "Wen Ding",
      "Yi Luan",
      "Sai Meher Karthik Duddu",
      "Gustavo Hernandez Abrego",
      "Weiqiang Shi",
      "Nithi Gupta",
      "Aditya Kusupati",
      "Prateek Jain",
      "Siddhartha Reddy Jonnalagadda",
      "Ming-Wei Chang",
      "Iftekhar Naim"
    ],
    "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-2403-20327v1",
        "title": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
        "url": "https://arxiv.org/abs/2403.20327v1"
      }
    ],
    "points": "65000",
    "published_grade": {
      "id": "lee-et-al-gecko-2024",
      "entry_id": "lee-et-al-gecko-2024",
      "band": "B",
      "modifier": "+",
      "target": {
        "record_id": "urn:uuid:35aee8ee-595d-4579-acac-ad0cc3a004d9",
        "version": 1,
        "content_hash": "sha256:35f4955a1d68192e17410de259ef04825fbfc9a60102a80f34a580d731457f10"
      },
      "what_changed": "Generated task descriptions and queries from web passages, retrieved alternative passages and used an LLM to relabel positives/hard negatives before contrastive embedding fine-tuning.",
      "scope": "Generated task descriptions and queries from web passages, retrieved alternative passages and used an LLM to relabel positives/hard negatives before contrastive embedding fine-tuning. The seed passage need not remain the best positive. Synthetic labels and few-shot relevance/query-likelihood judgments are not ground-truth guarantees. Teacher/data-generation and pretrained-backbone costs remain; dimension and leaderboard comparisons are source snapshots, not universal compression or current-rank claims.",
      "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:35aee8ee-595d-4579-acac-ad0cc3a004d9",
        "version": 1,
        "content_hash": "sha256:35f4955a1d68192e17410de259ef04825fbfc9a60102a80f34a580d731457f10"
      },
      "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": "Jinhyuk Lee",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Zhuyun Dai",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Xiaoqi Ren",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Blair Chen",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Daniel Cer",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Jeremy R. Cole",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Kai Hui",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Michael Boratko",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Rajvi Kapadia",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Wen Ding",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Yi Luan",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Sai Meher Karthik Duddu",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Gustavo Hernandez Abrego",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Weiqiang Shi",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Nithi Gupta",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Aditya Kusupati",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Prateek Jain",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Siddhartha Reddy Jonnalagadda",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Ming-Wei Chang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Iftekhar Naim",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
            "url": "https://arxiv.org/abs/2403.20327v1"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "The seed passage need not remain the best positive. Synthetic labels and few-shot relevance/query-likelihood judgments are not ground-truth guarantees. Teacher/data-generation and pretrained-backbone costs remain; dimension and leaderboard comparisons are source snapshots, not universal compression or current-rank claims.",
        "domain": "Text embedding training"
      },
      "sources": [
        {
          "citation": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
          "url": "https://arxiv.org/abs/2403.20327v1"
        }
      ],
      "statement": "# Gecko: task/query synthesis and passage relabeling for embedding distillation\n\nGenerated task descriptions and queries from web passages, retrieved alternative passages and used an LLM to relabel positives/hard negatives before contrastive embedding fine-tuning.\n\n## Scientific scope\n\nThe seed passage need not remain the best positive. Synthetic labels and few-shot relevance/query-likelihood judgments are not ground-truth guarantees. Teacher/data-generation and pretrained-backbone costs remain; dimension and leaderboard comparisons are source snapshots, not universal compression or current-rank claims.\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\nJinhyuk Lee; Zhuyun Dai; Xiaoqi Ren; Blair Chen; Daniel Cer; Jeremy R. Cole; Kai Hui; Michael Boratko; Rajvi Kapadia; Wen Ding; Yi Luan; Sai Meher Karthik Duddu; Gustavo Hernandez Abrego; Weiqiang Shi; Nithi Gupta; Aditya Kusupati; Prateek Jain; Siddhartha Reddy Jonnalagadda; Ming-Wei Chang; Iftekhar Naim\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- [Gecko: Versatile Text Embeddings Distilled from Large Language Models](https://arxiv.org/abs/2403.20327v1)\n",
      "summary": "Generated task descriptions and queries from web passages, retrieved alternative passages and used an LLM to relabel positives/hard negatives before contrastive embedding fine-tuning.",
      "title": "Gecko: task/query synthesis and passage relabeling for embedding distillation"
    },
    "record_id": "urn:uuid:35aee8ee-595d-4579-acac-ad0cc3a004d9",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:35aee8ee-595d-4579-acac-ad0cc3a004d9",
      "version": 1,
      "content_hash": "sha256:35f4955a1d68192e17410de259ef04825fbfc9a60102a80f34a580d731457f10"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:8d7b06f3422a64be1af3d5e8caeccdb789986205f453a5164cf4cdc3e4006bc9"
  },
  "target": {
    "record_id": "urn:uuid:35aee8ee-595d-4579-acac-ad0cc3a004d9",
    "version": 1,
    "content_hash": "sha256:35f4955a1d68192e17410de259ef04825fbfc9a60102a80f34a580d731457f10"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "lee-et-al-gecko-2024",
    "entry_id": "lee-et-al-gecko-2024",
    "revision": 1,
    "title": "Gecko: task/query synthesis and passage relabeling for embedding distillation",
    "context": "Text embedding training",
    "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": "Generated task descriptions and queries from web passages, retrieved alternative passages and used an LLM to relabel positives/hard negatives before contrastive embedding fine-tuning.",
    "scope": "Generated task descriptions and queries from web passages, retrieved alternative passages and used an LLM to relabel positives/hard negatives before contrastive embedding fine-tuning. The seed passage need not remain the best positive. Synthetic labels and few-shot relevance/query-likelihood judgments are not ground-truth guarantees. Teacher/data-generation and pretrained-backbone costs remain; dimension and leaderboard comparisons are source snapshots, not universal compression or current-rank claims.",
    "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. The seed passage need not remain the best positive. Synthetic labels and few-shot relevance/query-likelihood judgments are not ground-truth guarantees. Teacher/data-generation and pretrained-backbone costs remain; dimension and leaderboard comparisons are source snapshots, not universal compression or current-rank claims.",
    "references": [
      {
        "title": "Gecko: Versatile Text Embeddings Distilled from Large Language Models",
        "url": "https://arxiv.org/abs/2403.20327v1",
        "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:35aee8ee-595d-4579-acac-ad0cc3a004d9",
      "version": 1,
      "content_hash": "sha256:35f4955a1d68192e17410de259ef04825fbfc9a60102a80f34a580d731457f10"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "65000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}