---
dataset_info:
  features:
  - name: content
    dtype: string
  - name: markdown
    dtype: string
  - name: url
    dtype: string
  splits:
  - name: train
    num_bytes: 82642012
    num_examples: 2067
  download_size: 34862377
  dataset_size: 82642012
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

It is what it says on the tin, 2k websites that don't run javascript (aka: probably really old, and simple!). Filtering needs to be done to take out the defunct sites and error messages. I don't expect this is a hard task.