metadata
dataset_info:
features:
- name: task
dtype: string
- name: sub_task
dtype: string
- name: question
dtype: string
- name: context
dtype: float64
- name: answer
dtype: string
- name: rationale
dtype: string
- name: correct
dtype: bool
- name: instruction
dtype: string
- name: check
dtype: bool
- name: output
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 543370
num_examples: 400
download_size: 231568
dataset_size: 543370
configs:
- config_name: default
data_files:
- split: train
path: data/train-*