metadata
dataset_info:
features:
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- 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: 182861170.02451265
num_examples: 62018
download_size: 101304047
dataset_size: 182861170.02451265
configs:
- config_name: default
data_files:
- split: train
path: data/train-*