metadata
dataset_info:
features:
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 10628775.742184756
num_examples: 5194
- name: valid
num_bytes: 1023178.2578152441
num_examples: 500
download_size: 6181484
dataset_size: 11651954
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*