metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: weight
dtype: float64
- name: source_label_exact
sequence: string
splits:
- name: train
num_bytes: 1303533454
num_examples: 819313
download_size: 628163621
dataset_size: 1303533454
configs:
- config_name: default
data_files:
- split: train
path: data/train-*