---
dataset_info:
  features:
  - name: raw_content
    dtype: string
  - name: doc_id
    dtype: string
  - name: meta
    dtype: string
  - name: quality_signals
    dtype: string
  - name: ppl
    dtype: float64
  splits:
  - name: train
    num_bytes: 74782429843
    num_examples: 6391639
  download_size: 33367432066
  dataset_size: 74782429843
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---