metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: id
dtype: int64
- name: safety
dtype: int64
- name: acc_responses
dtype: string
- name: rej_responses
dtype: string
splits:
- name: train
num_bytes: 2221728
num_examples: 1256
download_size: 1037768
dataset_size: 2221728
configs:
- config_name: default
data_files:
- split: train
path: data/train-*