---
dataset_info:
  features:
  - name: question
    dtype: string
  - name: answer1
    dtype: string
  - name: answer2
    dtype: string
  - name: document
    dtype: string
  splits:
  - name: train
    num_bytes: 114807859
    num_examples: 32747
  download_size: 10198797
  dataset_size: 114807859
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: apache-2.0
---

Source dataset: <a href="https://huggingface.co/datasets/deepmind/narrativeqa">https://huggingface.co/datasets/deepmind/narrativeqa</a>