Datasets:

jplhughes2
/

alignment-faking-synthetic-chat-dataset-recall-0k-docs-8k-benign-2k-refusals

dataset_info:
  features:
    - name: prompt
      dtype: string
    - name: completion
      dtype: string
  splits:
    - name: train
      num_bytes: 12243803.374655647
      num_examples: 10000
  download_size: 10937596
  dataset_size: 12243803.374655647
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*