MultiTurn-Qwen3 / README.md
tuenguyen's picture
Upload dataset
690492a verified
metadata
dataset_info:
  features:
    - name: conversations
      list:
        - name: from
          dtype: string
        - name: info
          struct:
            - name: answer_content
              dtype: string
            - name: category
              dtype: string
            - name: ground_truth
              dtype: 'null'
            - name: instruction_constrain
              dtype: 'null'
            - name: ppl
              dtype: float64
            - name: source
              dtype: string
            - name: test_case
              dtype: 'null'
            - name: think_content
              dtype: string
            - name: verify_score
              dtype: float64
        - name: more
          list:
            - name: from
              dtype: string
            - name: info
              struct:
                - name: answer_content
                  dtype: string
                - name: coherence
                  dtype: 'null'
                - name: complexity
                  dtype: 'null'
                - name: correctness
                  dtype: 'null'
                - name: default_score
                  dtype: int64
                - name: format_satisfy
                  dtype: 'null'
                - name: ground_truth_verify_score
                  dtype: 'null'
                - name: helpfulness
                  dtype: 'null'
                - name: length
                  dtype: int64
                - name: md5
                  dtype: string
                - name: model_name
                  dtype: 'null'
                - name: ngram
                  dtype: 'null'
                - name: safety
                  dtype: 'null'
                - name: source
                  dtype: string
                - name: think_content
                  dtype: string
                - name: uuid
                  dtype: 'null'
                - name: verbosity
                  dtype: 'null'
            - name: value
              dtype: string
        - name: value
          dtype: string
    - name: system
      dtype: string
    - name: verify_score
      dtype: float64
    - name: question
      dtype: string
  splits:
    - name: train
      num_bytes: 2883428540
      num_examples: 143271
  download_size: 1544899041
  dataset_size: 2883428540
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*