metadata
dataset_info:
features:
- name: comparison_prompt
dtype: string
- name: messages
dtype: 'null'
- name: generation
dtype: 'null'
- name: model_name
dtype: 'null'
splits:
- name: train
num_bytes: 17706850
num_examples: 1083
download_size: 9845062
dataset_size: 17706850
configs:
- config_name: default
data_files:
- split: train
path: data/train-*