metadata
language:
- en
dataset_info:
features:
- name: _id
dtype: int64
- name: task
dtype: string
- name: tgt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 9448046
num_examples: 19823
download_size: 5458359
dataset_size: 9448046
configs:
- config_name: default
data_files:
- split: train
path: data/train-*