metadata
dataset_info:
features:
- name: chosen
dtype: string
- name: chosen_model
dtype: int64
- name: chosen_position
dtype: int64
- name: chosen_prompt
dtype: string
- name: id
dtype: int64
- name: reject
dtype: string
- name: reject_model
dtype: int64
- name: reject_prompt
dtype: string
splits:
- name: train
num_bytes: 15636939
num_examples: 963
download_size: 7959516
dataset_size: 15636939
configs:
- config_name: default
data_files:
- split: train
path: data/train-*