metadata
dataset_info:
features:
- name: conversations
struct:
- name: conversation
list:
- name: content
dtype: string
- name: role
dtype: string
- name: score
dtype: float64
- name: source
dtype: string
splits:
- name: train
num_bytes: 13359331
num_examples: 8789
download_size: 3735778
dataset_size: 13359331
configs:
- config_name: default
data_files:
- split: train
path: data/train-*