---
dataset_info:
  features:
  - name: model
    dtype: string
  - name: split
    dtype: string
  - name: label
    dtype:
      class_label:
        names:
          '0': safe
          '1': unsafe
  - name: case_id
    dtype: string
  - name: step_type
    dtype: string
  - name: step_index
    dtype: int32
  - name: image
    dtype: image
  splits:
  - name: train
    num_bytes: 4439339280
    num_examples: 63750
  download_size: 4522542270
  dataset_size: 4439339280
configs:
- config_name: default
  data_files:
  - split: train
    path: train/sd3/safe/train-*
---