metadata
dataset_info:
features:
- name: question
list:
- name: content
dtype: string
- name: role
dtype: string
- name: predicted_answer
dtype: string
- name: actual_answer
dtype: string
splits:
- name: train
num_bytes: 4397093
num_examples: 3200
download_size: 696407
dataset_size: 4397093
configs:
- config_name: default
data_files:
- split: train
path: data/train-*