metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: is_stem
dtype: string
splits:
- name: train
num_bytes: 95904488
num_examples: 26120
download_size: 50547303
dataset_size: 95904488
configs:
- config_name: default
data_files:
- split: train
path: data/train-*