---
dataset_info:
  features:
  - name: env
    dtype: string
  - name: prompt
    dtype: string
  - name: extra
    struct:
    - name: dataset_index
      dtype: int64
    - name: ground_truth
      dtype: string
    - name: inputs
      dtype: string
    - name: seed
      dtype: int64
    - name: task_id
      dtype: int64
    - name: transformed_code
      dtype: string
  - name: timestamp
    dtype: float64
  splits:
  - name: train
    num_bytes: 3554689531
    num_examples: 10000
  download_size: 3538901020
  dataset_size: 3554689531
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---