metadata
dataset_info:
features:
- name: id
dtype: string
- name: question
dtype: string
- name: system
dtype: string
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 2001009
num_examples: 920
download_size: 890932
dataset_size: 2001009
configs:
- config_name: default
data_files:
- split: train
path: data/train-*