---
dataset_info:
  features:
  - name: article
    dtype: string
  - name: article_sha1
    dtype: string
  - name: section_idx
    dtype: int32
  - name: section_title
    dtype: string
  - name: content
    dtype: string
  - name: question
    dtype: string
  - name: is_adversarial
    dtype: bool
  - name: input_ids
    sequence: int32
  - name: token_type_ids
    sequence: int8
  - name: attention_mask
    sequence: int8
  - name: label
    dtype: int64
  splits:
  - name: train
    num_bytes: 4042615849
    num_examples: 706498
  - name: validation
    num_bytes: 84704716
    num_examples: 15000
  download_size: 1555437560
  dataset_size: 4127320565
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---