{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "fabre-et-al-kairos-temporal-pretraining-2026",
    "title": "Kairos: temporal data ordering and factual-freshness evaluation",
    "summary": "Built a temporally grounded question benchmark and compared6B models pretrained on chronological versus shuffled versions of CommonCrawl data at a fixed token budget.",
    "boundary": "Chronological pretraining improved temporal freshness in the studied setting, not guaranteed continual learning, arbitrary knowledge updating or universal superiority. Released versus planned checkpoints must be checked separately at adoption.",
    "area": "Language-model data and evaluation",
    "topic": "Artificial intelligence",
    "year": 2026,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "fabre-et-al-kairos-temporal-pretraining-2026",
    "target": {
      "record_id": "urn:uuid:2c41af2c-9177-47dd-ac83-92962ac3238d",
      "version": 1,
      "content_hash": "sha256:bbd648fb485af3e2c97170a1395c7bbecdeed50f4a644d8168f80ba4c0d56c99"
    },
    "humans": [
      "Romain Fabre",
      "Hippolyte Pilchen",
      "Franck Signe Talla",
      "Patrick Perez",
      "Edouard Grave"
    ],
    "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-2605-22769v1",
        "title": "Understanding Data Temporality Impact on Large Language Models Pre-training",
        "url": "https://arxiv.org/abs/2605.22769v1"
      },
      {
        "id": "cur20261015-kyutai-publications-index",
        "title": "Kyutai scientific publications index",
        "url": "https://kyutai.org/papers/"
      }
    ],
    "points": "45000",
    "published_grade": {
      "id": "fabre-et-al-kairos-temporal-pretraining-2026",
      "entry_id": "fabre-et-al-kairos-temporal-pretraining-2026",
      "band": "B",
      "modifier": null,
      "target": {
        "record_id": "urn:uuid:2c41af2c-9177-47dd-ac83-92962ac3238d",
        "version": 1,
        "content_hash": "sha256:bbd648fb485af3e2c97170a1395c7bbecdeed50f4a644d8168f80ba4c0d56c99"
      },
      "what_changed": "Built a temporally grounded question benchmark and compared6B models pretrained on chronological versus shuffled versions of CommonCrawl data at a fixed token budget.",
      "scope": "Built a temporally grounded question benchmark and compared6B models pretrained on chronological versus shuffled versions of CommonCrawl data at a fixed token budget. Chronological pretraining improved temporal freshness in the studied setting, not guaranteed continual learning, arbitrary knowledge updating or universal superiority. Released versus planned checkpoints must be checked separately at adoption.",
      "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:2c41af2c-9177-47dd-ac83-92962ac3238d",
        "version": 1,
        "content_hash": "sha256:bbd648fb485af3e2c97170a1395c7bbecdeed50f4a644d8168f80ba4c0d56c99"
      },
      "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": "Romain Fabre",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Understanding Data Temporality Impact on Large Language Models Pre-training",
            "url": "https://arxiv.org/abs/2605.22769v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Hippolyte Pilchen",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Understanding Data Temporality Impact on Large Language Models Pre-training",
            "url": "https://arxiv.org/abs/2605.22769v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Franck Signe Talla",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Understanding Data Temporality Impact on Large Language Models Pre-training",
            "url": "https://arxiv.org/abs/2605.22769v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Patrick Perez",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Understanding Data Temporality Impact on Large Language Models Pre-training",
            "url": "https://arxiv.org/abs/2605.22769v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Edouard Grave",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Understanding Data Temporality Impact on Large Language Models Pre-training",
            "url": "https://arxiv.org/abs/2605.22769v1"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Understanding Data Temporality Impact on Large Language Models Pre-training",
            "url": "https://arxiv.org/abs/2605.22769v1"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-1",
          "kind": "source_reference",
          "source": {
            "citation": "Kyutai scientific publications index",
            "url": "https://kyutai.org/papers/"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "Chronological pretraining improved temporal freshness in the studied setting, not guaranteed continual learning, arbitrary knowledge updating or universal superiority. Released versus planned checkpoints must be checked separately at adoption.",
        "domain": "Language-model data and evaluation"
      },
      "sources": [
        {
          "citation": "Understanding Data Temporality Impact on Large Language Models Pre-training",
          "url": "https://arxiv.org/abs/2605.22769v1"
        },
        {
          "citation": "Kyutai scientific publications index",
          "url": "https://kyutai.org/papers/"
        }
      ],
      "statement": "# Kairos: temporal data ordering and factual-freshness evaluation\n\nBuilt a temporally grounded question benchmark and compared6B models pretrained on chronological versus shuffled versions of CommonCrawl data at a fixed token budget.\n\n## Scientific scope\n\nChronological pretraining improved temporal freshness in the studied setting, not guaranteed continual learning, arbitrary knowledge updating or universal superiority. Released versus planned checkpoints must be checked separately at adoption.\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\nRomain Fabre; Hippolyte Pilchen; Franck Signe Talla; Patrick Perez; Edouard Grave\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- [Understanding Data Temporality Impact on Large Language Models Pre-training](https://arxiv.org/abs/2605.22769v1)\n- [Kyutai scientific publications index](https://kyutai.org/papers/)\n",
      "summary": "Built a temporally grounded question benchmark and compared6B models pretrained on chronological versus shuffled versions of CommonCrawl data at a fixed token budget.",
      "title": "Kairos: temporal data ordering and factual-freshness evaluation"
    },
    "record_id": "urn:uuid:2c41af2c-9177-47dd-ac83-92962ac3238d",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:2c41af2c-9177-47dd-ac83-92962ac3238d",
      "version": 1,
      "content_hash": "sha256:bbd648fb485af3e2c97170a1395c7bbecdeed50f4a644d8168f80ba4c0d56c99"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:b84c2b9d97971a973dfbf3114eb9b163634dc40a353b82c32ff734e3c9dd228b"
  },
  "target": {
    "record_id": "urn:uuid:2c41af2c-9177-47dd-ac83-92962ac3238d",
    "version": 1,
    "content_hash": "sha256:bbd648fb485af3e2c97170a1395c7bbecdeed50f4a644d8168f80ba4c0d56c99"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "fabre-et-al-kairos-temporal-pretraining-2026",
    "entry_id": "fabre-et-al-kairos-temporal-pretraining-2026",
    "revision": 1,
    "title": "Kairos: temporal data ordering and factual-freshness evaluation",
    "context": "Language-model data and evaluation",
    "track": "Human-designed research method or resource",
    "band": "B",
    "modifier": null,
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-10-16",
    "evidence_cutoff": "2026-10-16",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Built a temporally grounded question benchmark and compared6B models pretrained on chronological versus shuffled versions of CommonCrawl data at a fixed token budget.",
    "scope": "Built a temporally grounded question benchmark and compared6B models pretrained on chronological versus shuffled versions of CommonCrawl data at a fixed token budget. Chronological pretraining improved temporal freshness in the studied setting, not guaranteed continual learning, arbitrary knowledge updating or universal superiority. Released versus planned checkpoints must be checked separately at adoption.",
    "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. Chronological pretraining improved temporal freshness in the studied setting, not guaranteed continual learning, arbitrary knowledge updating or universal superiority. Released versus planned checkpoints must be checked separately at adoption.",
    "references": [
      {
        "title": "Understanding Data Temporality Impact on Large Language Models Pre-training",
        "url": "https://arxiv.org/abs/2605.22769v1",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      },
      {
        "title": "Kyutai scientific publications index",
        "url": "https://kyutai.org/papers/",
        "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:2c41af2c-9177-47dd-ac83-92962ac3238d",
      "version": 1,
      "content_hash": "sha256:bbd648fb485af3e2c97170a1395c7bbecdeed50f4a644d8168f80ba4c0d56c99"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "45000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}