---
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: 37391220771.51984
    num_examples: 3195820
  download_size: 16340112876
  dataset_size: 37391220771.51984
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---