---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: perturbation_type
    dtype: string
  splits:
  - name: train
    num_bytes: 10651503
    num_examples: 25000
  download_size: 6133671
  dataset_size: 10651503
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---