metadata
dataset_info:
features:
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
- name: chosen_feature
sequence: float64
- name: rejected_feature
sequence: float64
splits:
- name: train
num_bytes: 542818143
num_examples: 7688
download_size: 110254404
dataset_size: 542818143
configs:
- config_name: default
data_files:
- split: train
path: data/train-*