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: 56937251
num_examples: 3022
download_size: 28704757
dataset_size: 56937251
configs:
- config_name: default
data_files:
- split: train
path: data/train-*