metadata
dataset_info:
features:
- name: dg_i
dtype: string
- name: dg_o
dtype: string
- name: ds_i
dtype: string
- name: ds_o
dtype: string
- name: dg_index
dtype: int64
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1037841630
num_examples: 129460
download_size: 201496033
dataset_size: 1037841630
configs:
- config_name: default
data_files:
- split: train
path: data/train-*