metadata
dataset_info:
features:
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: metadata
struct:
- name: found_rounds
sequence: int64
- name: items
sequence: int64
- name: task
dtype: string
splits:
- name: train
num_bytes: 67649774
num_examples: 43459
download_size: 28432063
dataset_size: 67649774
configs:
- config_name: default
data_files:
- split: train
path: data/train-*