---
dataset_info:
  features:
  - name: id
    dtype: string
  - name: question
    dtype: string
  - name: context
    sequence: string
  - name: response
    dtype: string
  - name: answer
    dtype: string
  splits:
  - name: train
    num_bytes: 358589480
    num_examples: 200500
  - name: test
    num_bytes: 877681
    num_examples: 500
  download_size: 206611798
  dataset_size: 359467161
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---


Dataset based on PubmedQA. Train and test split included. First 500 rows of train split are from labeled data, the last 200,000 are artifically labeled.