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: 15849390
num_examples: 2800
download_size: 4020898
dataset_size: 15849390
configs:
- config_name: default
data_files:
- split: train
path: data/train-*