{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "gunther-sturua-et-al-jina-embedding-family-2023",
    "title": "Jina embeddings v2/v3: long-context encoders and task-specific adapters",
    "summary": "Adapted bidirectional ALiBi and staged embedding training for longer documents in v2; v3 used an XLM-RoBERTa backbone with rotary positions, task-selected LoRA adapters and dimension-flexible representation training.",
    "boundary": "Version-specific architectures and author sets retained; v3 uses RoPE rather than v2's ALiBi. LoRA, rotary positions and Matryoshka learning are inherited tools. Shorter embedding dimensions trade quality for storage, despite the abstract's broad no-compromise wording. Longer accepted inputs are not certified lossless semantic compression.",
    "area": "Long-context text embeddings",
    "topic": "Artificial intelligence",
    "year": 2023,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "gunther-sturua-et-al-jina-embedding-family-2023",
    "target": {
      "record_id": "urn:uuid:88b0d857-44d0-4082-a05d-eebbdfed2974",
      "version": 1,
      "content_hash": "sha256:4284023d450dee049dc927f86b7aebfd4280fb0246fca5dca9370822e33ce000"
    },
    "humans": [
      "Michael Günther",
      "Jackmin Ong",
      "Isabelle Mohr",
      "Alaeddine Abdessalem",
      "Tanguy Abel",
      "Mohammad Kalim Akram",
      "Susana Guzman",
      "Georgios Mastrapas",
      "Saba Sturua",
      "Bo Wang",
      "Maximilian Werk",
      "Nan Wang",
      "Han Xiao",
      "Markus Krimmel",
      "Feng Wang",
      "Andreas Koukounas"
    ],
    "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-2310-19923v1",
        "title": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
        "url": "https://arxiv.org/abs/2310.19923v1"
      },
      {
        "id": "cur20261015-arxiv-2409-10173v1",
        "title": "jina-embeddings-v3: Multilingual Embeddings With Task LoRA",
        "url": "https://arxiv.org/abs/2409.10173v1"
      }
    ],
    "points": "75000",
    "published_grade": {
      "id": "gunther-sturua-et-al-jina-embedding-family-2023",
      "entry_id": "gunther-sturua-et-al-jina-embedding-family-2023",
      "band": "B",
      "modifier": "+",
      "target": {
        "record_id": "urn:uuid:88b0d857-44d0-4082-a05d-eebbdfed2974",
        "version": 1,
        "content_hash": "sha256:4284023d450dee049dc927f86b7aebfd4280fb0246fca5dca9370822e33ce000"
      },
      "what_changed": "Adapted bidirectional ALiBi and staged embedding training for longer documents in v2; v3 used an XLM-RoBERTa backbone with rotary positions, task-selected LoRA adapters and dimension-flexible representation training.",
      "scope": "Adapted bidirectional ALiBi and staged embedding training for longer documents in v2; v3 used an XLM-RoBERTa backbone with rotary positions, task-selected LoRA adapters and dimension-flexible representation training. Version-specific architectures and author sets retained; v3 uses RoPE rather than v2's ALiBi. LoRA, rotary positions and Matryoshka learning are inherited tools. Shorter embedding dimensions trade quality for storage, despite the abstract's broad no-compromise wording. Longer accepted inputs are not certified lossless semantic compression.",
      "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:88b0d857-44d0-4082-a05d-eebbdfed2974",
        "version": 1,
        "content_hash": "sha256:4284023d450dee049dc927f86b7aebfd4280fb0246fca5dca9370822e33ce000"
      },
      "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": "Michael Günther",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Jackmin Ong",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Isabelle Mohr",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Alaeddine Abdessalem",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Tanguy Abel",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Mohammad Kalim Akram",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Susana Guzman",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Georgios Mastrapas",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Saba Sturua",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Bo Wang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Maximilian Werk",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Nan Wang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Han Xiao",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Markus Krimmel",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "jina-embeddings-v3: Multilingual Embeddings With Task LoRA",
            "url": "https://arxiv.org/abs/2409.10173v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Feng Wang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "jina-embeddings-v3: Multilingual Embeddings With Task LoRA",
            "url": "https://arxiv.org/abs/2409.10173v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Andreas Koukounas",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "jina-embeddings-v3: Multilingual Embeddings With Task LoRA",
            "url": "https://arxiv.org/abs/2409.10173v1"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
            "url": "https://arxiv.org/abs/2310.19923v1"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-1",
          "kind": "source_reference",
          "source": {
            "citation": "jina-embeddings-v3: Multilingual Embeddings With Task LoRA",
            "url": "https://arxiv.org/abs/2409.10173v1"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "Version-specific architectures and author sets retained; v3 uses RoPE rather than v2's ALiBi. LoRA, rotary positions and Matryoshka learning are inherited tools. Shorter embedding dimensions trade quality for storage, despite the abstract's broad no-compromise wording. Longer accepted inputs are not certified lossless semantic compression.",
        "domain": "Long-context text embeddings"
      },
      "sources": [
        {
          "citation": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
          "url": "https://arxiv.org/abs/2310.19923v1"
        },
        {
          "citation": "jina-embeddings-v3: Multilingual Embeddings With Task LoRA",
          "url": "https://arxiv.org/abs/2409.10173v1"
        }
      ],
      "statement": "# Jina embeddings v2/v3: long-context encoders and task-specific adapters\n\nAdapted bidirectional ALiBi and staged embedding training for longer documents in v2; v3 used an XLM-RoBERTa backbone with rotary positions, task-selected LoRA adapters and dimension-flexible representation training.\n\n## Scientific scope\n\nVersion-specific architectures and author sets retained; v3 uses RoPE rather than v2's ALiBi. LoRA, rotary positions and Matryoshka learning are inherited tools. Shorter embedding dimensions trade quality for storage, despite the abstract's broad no-compromise wording. Longer accepted inputs are not certified lossless semantic compression.\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\nMichael Günther; Jackmin Ong; Isabelle Mohr; Alaeddine Abdessalem; Tanguy Abel; Mohammad Kalim Akram; Susana Guzman; Georgios Mastrapas; Saba Sturua; Bo Wang; Maximilian Werk; Nan Wang; Han Xiao; Markus Krimmel; Feng Wang; Andreas Koukounas\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- [Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents](https://arxiv.org/abs/2310.19923v1)\n- [jina-embeddings-v3: Multilingual Embeddings With Task LoRA](https://arxiv.org/abs/2409.10173v1)\n",
      "summary": "Adapted bidirectional ALiBi and staged embedding training for longer documents in v2; v3 used an XLM-RoBERTa backbone with rotary positions, task-selected LoRA adapters and dimension-flexible representation training.",
      "title": "Jina embeddings v2/v3: long-context encoders and task-specific adapters"
    },
    "record_id": "urn:uuid:88b0d857-44d0-4082-a05d-eebbdfed2974",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:88b0d857-44d0-4082-a05d-eebbdfed2974",
      "version": 1,
      "content_hash": "sha256:4284023d450dee049dc927f86b7aebfd4280fb0246fca5dca9370822e33ce000"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:d7dbc6075c900990ce122b9add700eaf44293097c407ed0f475af00b6143d962"
  },
  "target": {
    "record_id": "urn:uuid:88b0d857-44d0-4082-a05d-eebbdfed2974",
    "version": 1,
    "content_hash": "sha256:4284023d450dee049dc927f86b7aebfd4280fb0246fca5dca9370822e33ce000"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "gunther-sturua-et-al-jina-embedding-family-2023",
    "entry_id": "gunther-sturua-et-al-jina-embedding-family-2023",
    "revision": 1,
    "title": "Jina embeddings v2/v3: long-context encoders and task-specific adapters",
    "context": "Long-context text embeddings",
    "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": "Adapted bidirectional ALiBi and staged embedding training for longer documents in v2; v3 used an XLM-RoBERTa backbone with rotary positions, task-selected LoRA adapters and dimension-flexible representation training.",
    "scope": "Adapted bidirectional ALiBi and staged embedding training for longer documents in v2; v3 used an XLM-RoBERTa backbone with rotary positions, task-selected LoRA adapters and dimension-flexible representation training. Version-specific architectures and author sets retained; v3 uses RoPE rather than v2's ALiBi. LoRA, rotary positions and Matryoshka learning are inherited tools. Shorter embedding dimensions trade quality for storage, despite the abstract's broad no-compromise wording. Longer accepted inputs are not certified lossless semantic compression.",
    "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. Version-specific architectures and author sets retained; v3 uses RoPE rather than v2's ALiBi. LoRA, rotary positions and Matryoshka learning are inherited tools. Shorter embedding dimensions trade quality for storage, despite the abstract's broad no-compromise wording. Longer accepted inputs are not certified lossless semantic compression.",
    "references": [
      {
        "title": "Jina Embeddings 2: 8192-Token General-Purpose Text Embeddings for Long Documents",
        "url": "https://arxiv.org/abs/2310.19923v1",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      },
      {
        "title": "jina-embeddings-v3: Multilingual Embeddings With Task LoRA",
        "url": "https://arxiv.org/abs/2409.10173v1",
        "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:88b0d857-44d0-4082-a05d-eebbdfed2974",
      "version": 1,
      "content_hash": "sha256:4284023d450dee049dc927f86b7aebfd4280fb0246fca5dca9370822e33ce000"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "75000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}