metadata
dataset_info:
features:
- name: extra_info
struct:
- name: index
dtype: int64
- name: split
dtype: string
- name: groundtruth
dtype: string
- name: split
dtype: string
- name: messages
list:
- name: content
list:
- name: content
dtype: string
- name: role
dtype: string
- name: role
dtype: string
- name: dataset
dtype: string
- name: data_source
dtype: string
splits:
- name: train
num_bytes: 3702166753
num_examples: 482585
download_size: 1797096343
dataset_size: 3702166753
configs:
- config_name: default
data_files:
- split: train
path: data/train-*