metadata
dataset_info:
features:
- name: clean
dtype: string
- name: erroneous
dtype: string
- name: instruction
dtype: string
splits:
- name: train
num_bytes: 6059741586
num_examples: 4099711
- name: validation
num_bytes: 30505467
num_examples: 21205
- name: test
num_bytes: 23230217
num_examples: 16096
download_size: 1938604875
dataset_size: 6113477270
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*