---
dataset_info:
  features:
  - name: clf_label
    dtype:
      class_label:
        names:
          '0': Benign
          '1': Harmful
  - name: proxy_clf_label
    dtype:
      class_label:
        names:
          '0': Benign
          '1': Harmful
  - name: instructions
    dtype: string
  - name: content
    list: string
  - name: answer_prompt
    dtype: string
  - name: gen_target
    dtype: string
  - name: proxy_gen_target
    dtype: string
  splits:
  - name: train
    num_bytes: 82052
    num_examples: 313
  download_size: 52218
  dataset_size: 82052
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---