---
dataset_info:
  features:
  - name: task_id
    dtype: large_string
  - name: entry_point
    dtype: large_string
  - name: prompt
    dtype: large_string
  - name: variant_type
    dtype: large_string
  - name: is_adversarial
    dtype: bool
  - name: seed
    dtype: int64
  - name: transformation_order
    list: string
  - name: reference_solution
    dtype: large_string
  splits:
  - name: train
    num_bytes: 5032171
    num_examples: 5166
  download_size: 566545
  dataset_size: 5032171
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---