metadata
language:
- ja
license: cc-by-sa-4.0
task_categories:
- question-answering
dataset_info:
features:
- name: qid
dtype: string
- name: competition
dtype: string
- name: timestamp
dtype: string
- name: section
dtype: string
- name: number
dtype: int64
- name: original_question
dtype: string
- name: original_answer
dtype: string
- name: original_additional_info
dtype: string
- name: question
dtype: string
- name: answers
sequence: string
- name: original_answers
sequence: string
splits:
- name: validation
num_bytes: 453768
num_examples: 1000
download_size: 260118
dataset_size: 453768
configs:
- config_name: default
data_files:
- split: validation
path: data/validation-*
AIO with extended answers
AIO (AI王) is a Japanese quiz dataset. This repository contains the validation set of Version 2.0, augmented with manually annotated valid answers for each question.
Source: AI王 〜クイズAI日本一決定戦〜
Data Fields
qid
(str
): Unique identifier for each entry in the dataset.competition
(str
): The name of the competition that the data was first created for.timestamp
(str
): The timestamp when the data was created.section
(str
): The split the data belongs to.original_question
(str
): The original question before any processing or modification.original_answer
(str
): The original answer before any processing or modification.original_additional_info
(str
): Any additional information provided with the original data.question
(str
): The question used for the task. (NFKC 正規化済み)answers
(str
): The answer(s) of the question. This included additional answers we manually added to the original dataset.original_answers
(str
): The original "answers" of the original aio dataset.
Licensing Information
Creative Commons Attribution Share Alike 4.0 International
Note that the dataset includes a training set containing both copyright-protected data and data licensed under CC BY-SA 4.0, which are not included in this repository. The validation/test set is licensed under CC BY-SA 4.0.