metadata
dataset_info:
features:
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: og_dataset
dtype: string
- name: chosen_labeled
dtype: string
- name: rejected_labeled
dtype: string
- name: num_chosen_labels
dtype: int64
- name: num_rejected_labels
dtype: int64
splits:
- name: train
num_bytes: 23482845
num_examples: 5123
download_size: 12073739
dataset_size: 23482845
configs:
- config_name: default
data_files:
- split: train
path: data/train-*