metadata
dataset_info:
features:
- name: instruction
dtype: 'null'
- name: output
dtype: string
- name: origin
dtype: large_string
- name: run_id
dtype: large_string
- name: finish_reason
dtype: large_string
- name: out_tok
dtype: int64
splits:
- name: train
num_bytes: 92447004
num_examples: 12738
download_size: 50904612
dataset_size: 92447004
configs:
- config_name: default
data_files:
- split: train
path: data/train-*