README.md

metadata

dataset_info:
  features:
    - name: model_id
      dtype: string
    - name: dataset_id
      dtype: string
    - name: columns
      list: string
    - name: seed
      dtype: int64
    - name: sample_idx
      dtype: int64
    - name: sentence_prefix
      dtype: string
    - name: predicted_token
      dtype: string
    - name: actual_token
      dtype: string
    - name: probability
      dtype: float64
    - name: num_tokens
      dtype: int32
  splits:
    - name: train
      num_bytes: 1219194.5
      num_examples: 5000
  download_size: 484866
  dataset_size: 1219194.5
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*