metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: response
list:
- name: content
dtype: string
- name: role
dtype: string
- name: score
dtype: float64
splits:
- name: train_prefs
num_bytes: 320149736
num_examples: 122270
download_size: 126320171
dataset_size: 320149736
configs:
- config_name: default
data_files:
- split: train_prefs
path: data/train_prefs-*