metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: weight
dtype: float64
splits:
- name: train
num_bytes: 181795884
num_examples: 100000
download_size: 97226388
dataset_size: 181795884
configs:
- config_name: default
data_files:
- split: train
path: data/train-*