---
dataset_info:
  features:
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: source
    dtype: string
  - name: is_clean
    dtype: bool
  - name: applied_edits
    dtype: int64
  splits:
  - name: train
    num_bytes: 5645784
    num_examples: 9900
  - name: validation
    num_bytes: 55786
    num_examples: 100
  download_size: 1577653
  dataset_size: 5701570
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---