metadata
dataset_info:
features:
- name: instruction
dtype: 'null'
- name: output
dtype: string
- name: origin
dtype: string
- name: run_id
dtype: string
- name: finish_reason
dtype: string
- name: out_tok
dtype: int64
splits:
- name: train
num_bytes: 90606160
num_examples: 12499
download_size: 49983184
dataset_size: 90606160
configs:
- config_name: default
data_files:
- split: train
path: data/train-*