metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: response_A
dtype: string
- name: response_B
dtype: string
- name: reasoning
dtype: string
- name: preferred
dtype: string
splits:
- name: train
num_bytes: 475365
num_examples: 200
download_size: 301526
dataset_size: 475365
configs:
- config_name: default
data_files:
- split: train
path: data/train-*