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: 45763323
num_examples: 31513
- name: test
num_bytes: 5144532
num_examples: 3492
download_size: 28478873
dataset_size: 50907855
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*