{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "lai-et-al-ds1000-2022",
    "title": "DS-1000: data-science programming with execution and API constraints",
    "summary": "Curated and perturbed Stack Overflow problems across seven Python libraries, supplied executable contexts and combined functional tests with task-specific surface-form constraints.",
    "boundary": "The thousand tasks derive from a smaller set of source questions, not a thousand independent original posts. The evaluation environment fixes library versions. Perturbation and red teaming reduce memorization risk and observed false acceptance; they do not eliminate contamination or prove correctness.",
    "area": "Code-generation benchmarks",
    "topic": "Artificial intelligence",
    "year": 2022,
    "lane": "HUMAN",
    "result": "BENCHMARK",
    "contribution_kind": "EVALUATION_RESOURCE_DESIGN",
    "package_id": "lai-et-al-ds1000-2022",
    "target": {
      "record_id": "urn:uuid:b2ac8b80-b894-4ef3-abe8-22bd40609ade",
      "version": 1,
      "content_hash": "sha256:c59f8afe4256f44034a7dc99763e999a29f52718b7bb7ba05832fbdae154eedf"
    },
    "humans": [
      "Yuhang Lai",
      "Chengxi Li",
      "Yiming Wang",
      "Tianyi Zhang",
      "Ruiqi Zhong",
      "Luke Zettlemoyer",
      "Scott Wen-tau Yih",
      "Daniel Fried",
      "Sida Wang",
      "Tao Yu"
    ],
    "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-2211-11501v1",
        "title": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
        "url": "https://arxiv.org/abs/2211.11501v1"
      }
    ],
    "points": "115000",
    "published_grade": {
      "id": "lai-et-al-ds1000-2022",
      "entry_id": "lai-et-al-ds1000-2022",
      "band": "A",
      "modifier": null,
      "target": {
        "record_id": "urn:uuid:b2ac8b80-b894-4ef3-abe8-22bd40609ade",
        "version": 1,
        "content_hash": "sha256:c59f8afe4256f44034a7dc99763e999a29f52718b7bb7ba05832fbdae154eedf"
      },
      "what_changed": "Curated and perturbed Stack Overflow problems across seven Python libraries, supplied executable contexts and combined functional tests with task-specific surface-form constraints.",
      "scope": "Curated and perturbed Stack Overflow problems across seven Python libraries, supplied executable contexts and combined functional tests with task-specific surface-form constraints. The thousand tasks derive from a smaller set of source questions, not a thousand independent original posts. The evaluation environment fixes library versions. Perturbation and red teaming reduce memorization risk and observed false acceptance; they do not eliminate contamination or prove correctness.",
      "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:b2ac8b80-b894-4ef3-abe8-22bd40609ade",
        "version": 1,
        "content_hash": "sha256:c59f8afe4256f44034a7dc99763e999a29f52718b7bb7ba05832fbdae154eedf"
      },
      "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": "Yuhang Lai",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Chengxi Li",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Yiming Wang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Tianyi Zhang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Ruiqi Zhong",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Luke Zettlemoyer",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Scott Wen-tau Yih",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Daniel Fried",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Sida Wang",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        },
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Tao Yu",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
            "url": "https://arxiv.org/abs/2211.11501v1"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "The thousand tasks derive from a smaller set of source questions, not a thousand independent original posts. The evaluation environment fixes library versions. Perturbation and red teaming reduce memorization risk and observed false acceptance; they do not eliminate contamination or prove correctness.",
        "domain": "Code-generation benchmarks"
      },
      "sources": [
        {
          "citation": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
          "url": "https://arxiv.org/abs/2211.11501v1"
        }
      ],
      "statement": "# DS-1000: data-science programming with execution and API constraints\n\nCurated and perturbed Stack Overflow problems across seven Python libraries, supplied executable contexts and combined functional tests with task-specific surface-form constraints.\n\n## Scientific scope\n\nThe thousand tasks derive from a smaller set of source questions, not a thousand independent original posts. The evaluation environment fixes library versions. Perturbation and red teaming reduce memorization risk and observed false acceptance; they do not eliminate contamination or prove correctness.\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\nYuhang Lai; Chengxi Li; Yiming Wang; Tianyi Zhang; Ruiqi Zhong; Luke Zettlemoyer; Scott Wen-tau Yih; Daniel Fried; Sida Wang; Tao Yu\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- [DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation](https://arxiv.org/abs/2211.11501v1)\n",
      "summary": "Curated and perturbed Stack Overflow problems across seven Python libraries, supplied executable contexts and combined functional tests with task-specific surface-form constraints.",
      "title": "DS-1000: data-science programming with execution and API constraints"
    },
    "record_id": "urn:uuid:b2ac8b80-b894-4ef3-abe8-22bd40609ade",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:b2ac8b80-b894-4ef3-abe8-22bd40609ade",
      "version": 1,
      "content_hash": "sha256:c59f8afe4256f44034a7dc99763e999a29f52718b7bb7ba05832fbdae154eedf"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:589619bff00c2d3cf01be019340f6026d0b7f667c5750836c7caf94cf44e750d"
  },
  "target": {
    "record_id": "urn:uuid:b2ac8b80-b894-4ef3-abe8-22bd40609ade",
    "version": 1,
    "content_hash": "sha256:c59f8afe4256f44034a7dc99763e999a29f52718b7bb7ba05832fbdae154eedf"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "lai-et-al-ds1000-2022",
    "entry_id": "lai-et-al-ds1000-2022",
    "revision": 1,
    "title": "DS-1000: data-science programming with execution and API constraints",
    "context": "Code-generation benchmarks",
    "track": "Evaluation resource design",
    "band": "A",
    "modifier": null,
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-10-16",
    "evidence_cutoff": "2026-10-16",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Curated and perturbed Stack Overflow problems across seven Python libraries, supplied executable contexts and combined functional tests with task-specific surface-form constraints.",
    "scope": "Curated and perturbed Stack Overflow problems across seven Python libraries, supplied executable contexts and combined functional tests with task-specific surface-form constraints. The thousand tasks derive from a smaller set of source questions, not a thousand independent original posts. The evaluation environment fixes library versions. Perturbation and red teaming reduce memorization risk and observed false acceptance; they do not eliminate contamination or prove correctness.",
    "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 thousand tasks derive from a smaller set of source questions, not a thousand independent original posts. The evaluation environment fixes library versions. Perturbation and red teaming reduce memorization risk and observed false acceptance; they do not eliminate contamination or prove correctness.",
    "references": [
      {
        "title": "DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation",
        "url": "https://arxiv.org/abs/2211.11501v1",
        "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:b2ac8b80-b894-4ef3-abe8-22bd40609ade",
      "version": 1,
      "content_hash": "sha256:c59f8afe4256f44034a7dc99763e999a29f52718b7bb7ba05832fbdae154eedf"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "115000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}