metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: response_1
dtype: string
- name: response_2
dtype: string
- name: preferred_response
dtype: string
splits:
- name: train
num_bytes: 16989
num_examples: 10
download_size: 14081
dataset_size: 16989
configs:
- config_name: default
data_files:
- split: train
path: data/train-*