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: query
dtype: string
- name: length
dtype: int64
- name: query_token
sequence: int64
- name: chosen_text
dtype: string
- name: rejected_text
dtype: string
- name: query_chosen_token
sequence: int64
- name: query_rejected_token
sequence: int64
splits:
- name: train
num_bytes: 722446192
num_examples: 5524
- name: test
num_bytes: 12121412
num_examples: 102
download_size: 165286165
dataset_size: 734567604
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*