metadata
dataset_info:
features:
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 45574111
num_examples: 10162
- name: validation
num_bytes: 2319781
num_examples: 519
download_size: 23288200
dataset_size: 47893892
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*