---
dataset_info:
  features:
  - name: formatted_input
    dtype: string
  - name: correct_answer
    dtype: string
  - name: wrong_answer
    dtype: string
  - name: sent1
    dtype: string
  - name: intervention_insertion_pairs
    list:
    - name: insert
      dtype: string
    - name: interventioned_sentence
      dtype: string
  splits:
  - name: train
    num_bytes: 4664942
    num_examples: 10000
  download_size: 1301641
  dataset_size: 4664942
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---