metadata
dataset_info:
features:
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: question
struct:
- name: content
dtype: string
- name: role
dtype: string
- name: model_chosen
dtype: string
splits:
- name: train
num_bytes: 224707388
num_examples: 100000
download_size: 127458856
dataset_size: 224707388
configs:
- config_name: default
data_files:
- split: train
path: data/train-*