metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: token_length
dtype: int64
splits:
- name: train
num_bytes: 1535572543
num_examples: 100000
download_size: 676274539
dataset_size: 1535572543
configs:
- config_name: default
data_files:
- split: train
path: data/train-*