truthful_benchmark / README.md
ThWu's picture
Upload dataset
ee1eac1 verified
|
raw
history blame
476 Bytes
metadata
dataset_info:
  features:
    - name: question_id
      dtype: int64
    - name: prompt
      dtype: string
    - name: response_a
      dtype: string
    - name: response_b
      dtype: string
    - name: response_c
      dtype: string
    - name: ranked_responses
      sequence: string
  splits:
    - name: train
      num_bytes: 315878
      num_examples: 817
  download_size: 179861
  dataset_size: 315878
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*