metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: chosen_pref
dtype: string
- name: rejected_pref
dtype: string
- name: split_suffix
dtype: string
splits:
- name: train
num_bytes: 152828242
num_examples: 17788
- name: validation
num_bytes: 19388435
num_examples: 2280
download_size: 41371813
dataset_size: 172216677
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*