metadata
dataset_info:
features:
- name: id
dtype: string
- name: title
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: answers
struct:
- name: answer_start
sequence: int64
- name: text
sequence: string
- name: rejected
dtype: string
- name: chosen
dtype: string
splits:
- name: train
num_bytes: 9108
num_examples: 10
download_size: 11292
dataset_size: 9108
configs:
- config_name: default
data_files:
- split: train
path: data/train-*