README.md

metadata

dataset_info:
  features:
    - name: body
      dtype: string
    - name: comment_score
      dtype: int64
    - name: prompt
      dtype: string
    - name: post_score
      dtype: int64
  splits:
    - name: train
      num_bytes: 244506795.8945573
      num_examples: 66332
  download_size: 168000074
  dataset_size: 244506795.8945573
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: mit

Data from real humans, courtesy of https://reddit.com/r/WritingPrompts