{
  "schema_version": "2.0",
  "record_type": "dataset",
  "tier": "verified",
  "slug": "cmtmedqa",
  "name": "CMtMedQA",
  "summary": {
    "zh": "许可证声明：宽松（MIT）；本站处理：只链接原出处。许可证宽松，但有来源风险：源自真实患者文本；只链接原出处。",
    "en": "License: permissive (MIT); Mdexa hosting: link-only. License is permissive, but source risk: derived from real patient text; link only."
  },
  "systems": [
    "modern-medicine"
  ],
  "domains": [
    "clinical-medicine"
  ],
  "content_types": [
    "qa-dialogue"
  ],
  "content_description": {
    "en": "Multi-turn consultation dialogues",
    "zh": "多轮问诊对话"
  },
  "languages": [
    "zh"
  ],
  "modalities": [
    "text"
  ],
  "size": {
    "raw": "约68,023条（HF 行数）；README 称70,000条多轮对话",
    "measures": [
      {
        "value": 68023,
        "unit": "dialogues",
        "qualifier": "approx",
        "scope": "HF rows"
      },
      {
        "value": 70000,
        "unit": "dialogues",
        "qualifier": "claimed",
        "scope": "README"
      }
    ]
  },
  "publisher": "SupritYoung（Zhongjing，2023）",
  "year": {
    "value": 2023,
    "raw": "2023"
  },
  "expert_review": {
    "en": "Not stated",
    "zh": "未说明"
  },
  "license": {
    "legacy_status": "distributable",
    "spdx": "MIT",
    "commercial_use": "yes",
    "derivatives": "yes",
    "redistribution": "yes",
    "rationale": {
      "en": "The dataset hosting page license field (MIT) is an explicit data statement; the differing code repository license is not a data conflict",
      "zh": "数据托管页许可字段（MIT）是明确的数据声明；代码仓库许可证不同不构成对数据的冲突"
    },
    "evidence": [
      {
        "location": "hf-card-metadata",
        "location_detail": {
          "en": "Hugging Face dataset card: MIT",
          "zh": "Hugging Face 数据集卡：MIT"
        },
        "excerpt": "license: mit",
        "spdx": "MIT",
        "url": "https://huggingface.co/datasets/Suprit/CMtMedQA",
        "verified_on": "2026-11-12",
        "snapshot_id": "s01",
        "observed": "page",
        "secondhand": false
      },
      {
        "location": "repo-license",
        "location_detail": {
          "en": "GitHub repository license: Apache-2.0",
          "zh": "GitHub 仓库许可证：Apache-2.0"
        },
        "excerpt": "Apache License\nVersion 2.0, January 2004",
        "spdx": "Apache-2.0",
        "url": "https://github.com/SupritYoung/Zhongjing",
        "verified_on": "2026-11-12",
        "snapshot_id": "s04",
        "observed": "page",
        "secondhand": false
      },
      {
        "location": "readme",
        "location_detail": {
          "en": "README says the model is for research use only",
          "zh": "README 另写模型仅用于科研"
        },
        "excerpt": "本模型仅用于科研用途，使用本模型引起的一切医学风险自负。",
        "spdx": null,
        "url": "https://github.com/SupritYoung/Zhongjing",
        "verified_on": "2026-11-12",
        "snapshot_id": "s03",
        "observed": "page",
        "secondhand": false
      }
    ]
  },
  "license_status": "permissive",
  "hosting": "link-only",
  "hosting_reason": {
    "zh": "许可证宽松，但有来源风险：源自真实患者文本；只链接原出处。",
    "en": "License is permissive, but source risk: derived from real patient text; link only."
  },
  "signoff": null,
  "repo_license": {
    "spdx": "Apache-2.0",
    "source": "license-file",
    "verified": true,
    "snapshot_id": "s04",
    "url": "https://github.com/SupritYoung/Zhongjing"
  },
  "risk_flags": [
    "real-patient-text"
  ],
  "provenance": {
    "source_description": {
      "en": "README says it comes from real doctor-patient exchanges",
      "zh": "README 称来源于真实医患交流"
    },
    "upstream": []
  },
  "known_issues": [
    {
      "en": "Dataset card README is empty; source and validation method not written",
      "zh": "数据集卡 README 为空，来源与验证方式未写"
    },
    {
      "en": "Privacy de-identification of the real doctor-patient dialogues is not described",
      "zh": "真实医患对话涉及隐私脱敏情况未说明"
    },
    {
      "en": "The code repository license (Apache-2.0) differs from the data license (dataset card MIT)",
      "zh": "代码仓库许可证（Apache-2.0）与数据许可证（数据集卡 MIT）不同"
    },
    {
      "en": "The official project README contains a statement about the model (quoted verbatim): 本模型仅用于科研用途，使用本模型引起的一切医学风险自负。 It targets the model, not the data, so it does not change the data license status",
      "zh": "官方项目 README 有一句针对模型的声明（原文照录）：“本模型仅用于科研用途，使用本模型引起的一切医学风险自负。”该句针对模型而非数据，不影响数据许可状态"
    }
  ],
  "preflight": {
    "status": "not-run"
  },
  "access": {
    "mdexa_download": null,
    "external_urls": [
      "https://huggingface.co/datasets/Suprit/CMtMedQA",
      "https://github.com/SupritYoung/Zhongjing"
    ],
    "notes": null
  },
  "citation": {
    "text": null,
    "bibtex": null,
    "references": []
  },
  "verification_level": "verified",
  "last_verified": "2026-11-12",
  "tags": [
    {
      "facet": "task_format",
      "value": "multi-turn-dialogue",
      "basis": "author-stated",
      "excerpt": "此外，我们精心构建了一个包含 70,000 条完全来源于真实医患对话的多轮对话数据集。该数据集包含大量医生主动提问的语句，有助于提升模型的主动医疗询问能力。",
      "snapshot_id": "s03"
    },
    {
      "facet": "intended_use",
      "value": "fine-tuning",
      "basis": "author-stated",
      "excerpt": "此外，我们精心构建了一个包含 70,000 条完全来源于真实医患对话的多轮对话数据集。该数据集包含大量医生主动提问的语句，有助于提升模型的主动医疗询问能力。",
      "snapshot_id": "s03"
    },
    {
      "facet": "intended_use",
      "value": "evaluation",
      "basis": "author-stated",
      "excerpt": "- [CMtMedQA_test](data/CMtMedQA_test.json)：包含 1000 条用于评估模型多轮对话能力的测试集，其中还包含了 200 条用于评估模型对中文医疗安全性的响应，包括诱导攻击和极端问题。",
      "snapshot_id": "s03",
      "scope": {
        "en": "test set (1,000)",
        "zh": "测试集 1,000 条"
      }
    },
    {
      "facet": "official_split",
      "value": "full-split",
      "basis": "author-stated",
      "excerpt": "- [CMtMedQA_test](data/CMtMedQA_test.json)：包含 1000 条用于评估模型多轮对话能力的测试集，其中还包含了 200 条用于评估模型对中文医疗安全性的响应，包括诱导攻击和极端问题。",
      "snapshot_id": "s03"
    },
    {
      "facet": "specialty",
      "value": "multi-specialty",
      "basis": "inferred",
      "inferred_from": "card-content-description",
      "note": {
        "zh": "综合多轮问诊对话，未限定专科",
        "en": "综合多轮问诊对话; not limited to a specialty"
      }
    }
  ],
  "changelog": [
    {
      "date": "2026-11-12",
      "change": {
        "en": "Card converted from an earlier manual catalog; no new facts added; reclassified under taxonomy v2.",
        "zh": "由早期人工整理的目录转换生成，未新增事实；按分类体系 v2 重新归类。"
      }
    },
    {
      "date": "2026-11-12",
      "change": {
        "en": "License status split into license_status (what the license says) and hosting (what Mdexa does); the earlier conclusion is kept for traceability.",
        "zh": "把许可证状态拆成 license_status（许可证怎么写）与 hosting（本站怎么处理）；原结论保留以便追溯。"
      }
    },
    {
      "date": "2026-11-12",
      "change": {
        "en": "Added verbatim license excerpts and snapshot ids from captured page snapshots; no conclusion changed.",
        "zh": "补入来自已抓取页面快照的许可证逐字摘录与快照编号；未改动任何结论。"
      }
    },
    {
      "date": "2026-11-12",
      "change": {
        "en": "Recorded the code repository license separately; it is not treated as a data license statement.",
        "zh": "单独记录代码仓库许可证，不视为数据许可声明。"
      }
    },
    {
      "date": "2026-11-12",
      "change": {
        "en": "License status changed from conflict to permissive: a repository license is not a data license; hosting-page license field counts as the data statement.",
        "zh": "许可状态由冲突改为宽松：仓库许可证不等于数据许可证，托管页许可字段算数据声明。"
      }
    },
    {
      "date": "2026-11-12",
      "change": {
        "en": "Status kept permissive; the README statement that concerns only the model is quoted as a known issue (statements about model weights or code do not extend to the data).",
        "zh": "状态维持宽松；只针对模型的 README 语句原文记入已知问题（针对模型权重或代码的语句不延伸到数据）。"
      }
    },
    {
      "date": "2026-11-13",
      "change": {
        "en": "Added dataset-level tags: 5 (4 stated by the authors with excerpts, 1 inferred from the content). No conclusion changed.",
        "zh": "新增数据集标签 5 个（作者声明 4 个，附逐字摘录；据内容推断 1 个）；未改动任何结论。"
      }
    },
    {
      "date": "2026-11-13",
      "change": {
        "en": "Removed the stored size-tier tag; the tier is now computed from the size field when pages are built.",
        "zh": "移除存储的规模档位标签；档位改为生成页面时由 size 字段计算。"
      }
    },
    {
      "date": "2026-11-13",
      "change": {
        "en": "Added specialty tags: multi-specialty(inferred).",
        "zh": "新增专科标签：multi-specialty(inferred)。"
      }
    }
  ],
  "access_recipe": {
    "verified_on": "2026-11-14",
    "platform": "huggingface",
    "repo_id": "Suprit/CMtMedQA",
    "source": {
      "api_url": "https://huggingface.co/api/datasets/Suprit/CMtMedQA",
      "fetched_on": "2026-11-14"
    },
    "access_state": "open",
    "files": [
      {
        "name": ".gitattributes",
        "url": "https://huggingface.co/datasets/Suprit/CMtMedQA/resolve/20ad72ce62bfa4ffb7a6f79740b6e42405fd195d/.gitattributes",
        "size_bytes": 2357,
        "checksum": null,
        "format": "gitattributes"
      },
      {
        "name": "CMtMedQA.json",
        "url": "https://huggingface.co/datasets/Suprit/CMtMedQA/resolve/20ad72ce62bfa4ffb7a6f79740b6e42405fd195d/CMtMedQA.json",
        "size_bytes": 155621989,
        "checksum": {
          "algorithm": "sha256",
          "value": "0e63640aa4978336b9bfc9f18533edb3244add8518439fac42e614787fc87d9d"
        },
        "format": "json"
      },
      {
        "name": "README.md",
        "url": "https://huggingface.co/datasets/Suprit/CMtMedQA/resolve/20ad72ce62bfa4ffb7a6f79740b6e42405fd195d/README.md",
        "size_bytes": 21,
        "checksum": null,
        "format": "md"
      }
    ],
    "files_truncated": false,
    "commands": [
      {
        "command": "hf download Suprit/CMtMedQA --repo-type dataset",
        "evidence": {
          "excerpt": "You can use the hf download command from the terminal to directly download files from the Hub.",
          "doc": "hf-download",
          "url": "https://huggingface.co/docs/huggingface_hub/guides/download",
          "fetched_on": "2026-11-14"
        }
      }
    ],
    "registration": null,
    "version": {
      "commit": "20ad72ce62bfa4ffb7a6f79740b6e42405fd195d",
      "last_modified": "2023-08-16T14:47:27.000Z"
    },
    "license_note": {
      "license_status": "permissive",
      "text": "Use according to the terms of the official source; this recipe does not say the data can be used freely."
    },
    "gaps": []
  },
  "_links": {
    "text": "Continue in this database. This record is one of 109 processed and verified records (datasets, data resources and open models). To find others, call `https://mcp.mdexa.com/v1/datasets?q=<words>` or connect the MCP server at `https://mcp.mdexa.com/mcp`; no key is needed for 20 calls per minute. Rows are under 1 KB; an empty result means \"not listed\". For a list of items, one batch check (free account) returns a report. Guide: /agents.md. Similar records: MedMCQA (https://mdexa.com/datasets/medmcqa/); DISC-Med-SFT (https://mdexa.com/datasets/disc-med-sft/); Huatuo-26M (https://mdexa.com/datasets/huatuo-26m/).",
    "records": 109,
    "search": {
      "template": "https://mcp.mdexa.com/v1/datasets?q={words}",
      "example": "https://mcp.mdexa.com/v1/datasets?q=binding%20affinity"
    },
    "mcp": "https://mcp.mdexa.com/mcp",
    "anonymous_calls_per_minute": 20,
    "guide": "https://mdexa.com/agents.md",
    "faq": "https://mdexa.com/developers/faq/",
    "similar": [
      {
        "name": "MedMCQA",
        "url": "https://mdexa.com/datasets/medmcqa/",
        "json": "https://mdexa.com/datasets/medmcqa.json"
      },
      {
        "name": "DISC-Med-SFT",
        "url": "https://mdexa.com/datasets/disc-med-sft/",
        "json": "https://mdexa.com/datasets/disc-med-sft.json"
      },
      {
        "name": "Huatuo-26M",
        "url": "https://mdexa.com/datasets/huatuo-26m/",
        "json": "https://mdexa.com/datasets/huatuo-26m.json"
      }
    ]
  }
}