metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 40934951.200845666
num_examples: 16602
- name: test
num_bytes: 4549149.799154334
num_examples: 1845
download_size: 21723636
dataset_size: 45484101
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*