README.md

metadata

dataset_info:
  - config_name: context
    features:
      - name: background
        dtype: string
      - name: questions
        list:
          - name: answer
            dtype: string
          - name: option
            struct:
              - name: A
                dtype: string
              - name: B
                dtype: string
              - name: C
                dtype: string
              - name: D
                dtype: string
              - name: E
                dtype: string
              - name: F
                dtype: string
          - name: question
            dtype: string
      - name: num_questions
        dtype: int64
      - name: exam_type
        dtype: string
      - name: exam_subject
        dtype: string
    splits:
      - name: train
        num_bytes: 1013026
        num_examples: 1221
      - name: val
        num_bytes: 3588
        num_examples: 2
      - name: test
        num_bytes: 3580
        num_examples: 6
    download_size: 616557
    dataset_size: 1020194
  - config_name: default
    features:
      - name: exam_type
        dtype: string
      - name: exam_class
        dtype: string
      - name: exam_subject
        dtype: string
      - name: question
        dtype: string
      - name: answer
        dtype: string
      - name: question_type
        dtype: string
      - name: option
        struct:
          - name: A
            dtype: string
          - name: B
            dtype: string
          - name: C
            dtype: string
          - name: D
            dtype: string
          - name: E
            dtype: string
          - name: F
            dtype: string
      - name: medical_terms
        list: string
    splits:
      - name: train
        num_bytes: 85316691
        num_examples: 269359
      - name: val
        num_bytes: 106698
        num_examples: 280
      - name: test
        num_bytes: 3548549
        num_examples: 11200
    download_size: 40457932
    dataset_size: 88971938
  - config_name: subdomain
    features:
      - name: medical_term
        dtype: string
      - name: questions
        list:
          - name: answer
            dtype: string
          - name: exam_subject
            dtype: string
          - name: exam_type
            dtype: string
          - name: option
            struct:
              - name: A
                dtype: string
              - name: B
                dtype: string
              - name: C
                dtype: string
              - name: D
                dtype: string
              - name: E
                dtype: string
              - name: F
                dtype: string
          - name: question
            dtype: string
      - name: num_questions
        dtype: int64
    splits:
      - name: train
        num_bytes: 12939714
        num_examples: 5674
      - name: val
        num_bytes: 35540
        num_examples: 21
      - name: test
        num_bytes: 1474959
        num_examples: 831
    download_size: 7248880
    dataset_size: 14450213
configs:
  - config_name: context
    data_files:
      - split: train
        path: context/train-*
      - split: val
        path: context/val-*
      - split: test
        path: context/test-*
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: val
        path: data/val-*
      - split: test
        path: data/test-*
  - config_name: subdomain
    data_files:
      - split: train
        path: subdomain/train-*
      - split: val
        path: subdomain/val-*
      - split: test
        path: subdomain/test-*

CMB-Exam-Grouped

This dataset contains medical exam questions with shared background context extracted and grouped together.

Features

• question_id: Unique identifier for each question
• background_id: ID for grouping questions that share the same background (-1 means no shared background)
• background: Shared context/background text extracted from questions
• question: The actual question (with background removed if applicable)
• option: Multiple choice options (A-F)
• answer: Correct answer
• exam_type, exam_class, exam_subject: Exam metadata

Statistics

• Train: 2,062 samples with shared backgrounds
• Val: 7 samples
• Test: 6 samples