metadata
dataset_info:
features:
- name: prompt
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: 499886
num_examples: 1024
- name: test
num_bytes: 124901
num_examples: 256
download_size: 116789
dataset_size: 624787
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*