metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: original_response
dtype: string
- name: explanation
dtype: string
- name: new_response
dtype: string
splits:
- name: train
num_bytes: 3421103
num_examples: 546
download_size: 1686318
dataset_size: 3421103
configs:
- config_name: default
data_files:
- split: train
path: data/train-*