sorenmulli/citizenship-test-da

---
dataset_info:
- config_name: default
features:
- name: question
dtype: string
- name: index
dtype: int64
- name: option-A
dtype: string
- name: option-B
dtype: string
- name: option-C
dtype: string
- name: correct
dtype: string
- name: origin
dtype: string
splits:
- name: train
num_bytes: 103251.0
num_examples: 605
download_size: 43667
dataset_size: 103251.0
- config_name: raw
features:
- name: question
dtype: string
- name: index
dtype: int64
- name: option-A
dtype: string
- name: option-B
dtype: string
- name: option-C
dtype: string
- name: correct
dtype: string
- name: origin
dtype: string
splits:
- name: train
num_bytes: 103906
num_examples: 605
download_size: 45297
dataset_size: 103906
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- config_name: raw
data_files:
- split: train
path: raw/train-*
---
# [WIP] Dataset Card for "citizenship-test-da"

*Please note that this dataset and dataset card both are works in progress. For now refer to the related [thesis](https://sorenmulli.github.io/thesis/thesis.pdf) for all details*

This dataset contains scraped questions and answers from Danish citizen tests (Danish: *indfødsretsprøver* og *medborgerskabsprøver*) from Juni 2019 to May 2023 from PDF's produced by ''Styrelsen for International Rekruttering og Integration'' (SIRI).

The dataset is released as an appendix to the thesis [''Are GLLMs Danoliterate? Benchmarking Generative NLP in Danish''](https://sorenmulli.github.io/thesis/thesis.pdf) and permission by SIRI for this specific purpose.

The PDF's are available on [SIRI's website](https://siri.dk/nyheder/?categorizations=9115).
The `default` configuration has been semi-automatically cleaned to remove PDF artifacts using the [Alvenir 3gram DSL language model](https://github.com/danspeech/danspeech/releases/tag/v0.02-alpha).
The examples were not deduplicated.

原始信息汇总

数据集概述

数据集配置

• default

  • 特征:
    • question: string
    • index: int64
    • option-A: string
    • option-B: string
    • option-C: string
    • correct: string
    • origin: string
  • 分割:
    • train
      • 字节数: 103251.0
      • 样本数: 605
  • 下载大小: 43667
  • 数据集大小: 103251.0

• raw

  • 特征:
    • question: string
    • index: int64
    • option-A: string
    • option-B: string
    • option-C: string
    • correct: string
    • origin: string
  • 分割:
    • train
      • 字节数: 103906
      • 样本数: 605
  • 下载大小: 45297
  • 数据集大小: 103906

数据文件

• default

  • 分割:
    • train: data/train-*

• raw

  • 分割:
    • train: raw/train-*