metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: input
dtype: string
- name: target_A
dtype: string
- name: target_B
dtype: string
splits:
- name: train
num_bytes: 4562580
num_examples: 5000
- name: eval
num_bytes: 451259
num_examples: 500
download_size: 2196260
dataset_size: 5013839
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: eval
path: data/eval-*