---
annotations_creators:
- found
language_creators:
- found
language:
- ar
- bg
- de
- es
- fr
- hr
- hu
- it
- lt
- mk
- pl
- pt
- sq
- sr
- tr
- vi
license:
- cc-by-sa-4.0
multilinguality:
- monolingual
- multilingual
size_categories:
- 10K<n<100K
- 1K<n<10K
- n<1K
source_datasets:
- original
task_categories:
- question-answering
task_ids:
- multiple-choice-qa
paperswithcode_id: exams
pretty_name: EXAMS
config_names:
- alignments
- crosslingual_bg
- crosslingual_hr
- crosslingual_hu
- crosslingual_it
- crosslingual_mk
- crosslingual_pl
- crosslingual_pt
- crosslingual_sq
- crosslingual_sr
- crosslingual_test
- crosslingual_tr
- crosslingual_vi
- crosslingual_with_para_bg
- crosslingual_with_para_hr
- crosslingual_with_para_hu
- crosslingual_with_para_it
- crosslingual_with_para_mk
- crosslingual_with_para_pl
- crosslingual_with_para_pt
- crosslingual_with_para_sq
- crosslingual_with_para_sr
- crosslingual_with_para_test
- crosslingual_with_para_tr
- crosslingual_with_para_vi
- multilingual
- multilingual_with_para
dataset_info:
- config_name: alignments
  features:
  - name: source_id
    dtype: string
  - name: target_id_list
    sequence: string
  splits:
  - name: full
    num_bytes: 1265256
    num_examples: 10834
  download_size: 184096
  dataset_size: 1265256
- config_name: crosslingual_bg
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 1077329
    num_examples: 2344
  - name: validation
    num_bytes: 281771
    num_examples: 593
  download_size: 514922
  dataset_size: 1359100
- config_name: crosslingual_hr
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 807104
    num_examples: 2341
  - name: validation
    num_bytes: 176594
    num_examples: 538
  download_size: 450090
  dataset_size: 983698
- config_name: crosslingual_hu
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 677535
    num_examples: 1731
  - name: validation
    num_bytes: 202012
    num_examples: 536
  download_size: 401455
  dataset_size: 879547
- config_name: crosslingual_it
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 399312
    num_examples: 1010
  - name: validation
    num_bytes: 93175
    num_examples: 246
  download_size: 226376
  dataset_size: 492487
- config_name: crosslingual_mk
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 825702
    num_examples: 1665
  - name: validation
    num_bytes: 204318
    num_examples: 410
  download_size: 394548
  dataset_size: 1030020
- config_name: crosslingual_pl
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 573410
    num_examples: 1577
  - name: validation
    num_bytes: 141633
    num_examples: 394
  download_size: 341925
  dataset_size: 715043
- config_name: crosslingual_pt
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 374798
    num_examples: 740
  - name: validation
    num_bytes: 87714
    num_examples: 184
  download_size: 208021
  dataset_size: 462512
- config_name: crosslingual_sq
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 423744
    num_examples: 1194
  - name: validation
    num_bytes: 110093
    num_examples: 311
  download_size: 247052
  dataset_size: 533837
- config_name: crosslingual_sr
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 649560
    num_examples: 1323
  - name: validation
    num_bytes: 145724
    num_examples: 314
  download_size: 327466
  dataset_size: 795284
- config_name: crosslingual_test
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: test
    num_bytes: 8402575
    num_examples: 19736
  download_size: 3438526
  dataset_size: 8402575
- config_name: crosslingual_tr
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 717599
    num_examples: 1571
  - name: validation
    num_bytes: 182730
    num_examples: 393
  download_size: 440914
  dataset_size: 900329
- config_name: crosslingual_vi
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 953167
    num_examples: 1955
  - name: validation
    num_bytes: 231976
    num_examples: 488
  download_size: 462940
  dataset_size: 1185143
- config_name: crosslingual_with_para_bg
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 47066808
    num_examples: 2344
  - name: validation
    num_bytes: 11916026
    num_examples: 593
  download_size: 15794611
  dataset_size: 58982834
- config_name: crosslingual_with_para_hr
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 24889604
    num_examples: 2341
  - name: validation
    num_bytes: 5695066
    num_examples: 538
  download_size: 9839452
  dataset_size: 30584670
- config_name: crosslingual_with_para_hu
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 19035663
    num_examples: 1731
  - name: validation
    num_bytes: 6043265
    num_examples: 536
  download_size: 9263625
  dataset_size: 25078928
- config_name: crosslingual_with_para_it
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 16409235
    num_examples: 1010
  - name: validation
    num_bytes: 4018329
    num_examples: 246
  download_size: 6907617
  dataset_size: 20427564
- config_name: crosslingual_with_para_mk
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 38445894
    num_examples: 1665
  - name: validation
    num_bytes: 9673574
    num_examples: 410
  download_size: 12878474
  dataset_size: 48119468
- config_name: crosslingual_with_para_pl
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 16373781
    num_examples: 1577
  - name: validation
    num_bytes: 4158832
    num_examples: 394
  download_size: 6539172
  dataset_size: 20532613
- config_name: crosslingual_with_para_pt
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 12185383
    num_examples: 740
  - name: validation
    num_bytes: 3093712
    num_examples: 184
  download_size: 4956969
  dataset_size: 15279095
- config_name: crosslingual_with_para_sq
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 17341277
    num_examples: 1194
  - name: validation
    num_bytes: 4449952
    num_examples: 311
  download_size: 7112236
  dataset_size: 21791229
- config_name: crosslingual_with_para_sr
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 24575845
    num_examples: 1323
  - name: validation
    num_bytes: 5772509
    num_examples: 314
  download_size: 8035415
  dataset_size: 30348354
- config_name: crosslingual_with_para_test
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: test
    num_bytes: 207974374
    num_examples: 13510
  download_size: 62878029
  dataset_size: 207974374
- config_name: crosslingual_with_para_tr
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 18597131
    num_examples: 1571
  - name: validation
    num_bytes: 4763097
    num_examples: 393
  download_size: 7346658
  dataset_size: 23360228
- config_name: crosslingual_with_para_vi
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 40882999
    num_examples: 1955
  - name: validation
    num_bytes: 10260374
    num_examples: 488
  download_size: 13028078
  dataset_size: 51143373
- config_name: multilingual
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 3381837
    num_examples: 7961
  - name: validation
    num_bytes: 1141687
    num_examples: 2672
  - name: test
    num_bytes: 5746781
    num_examples: 13510
  download_size: 4323915
  dataset_size: 10270305
- config_name: multilingual_with_para
  features:
  - name: id
    dtype: string
  - name: question
    struct:
    - name: stem
      dtype: string
    - name: choices
      sequence:
      - name: text
        dtype: string
      - name: label
        dtype: string
      - name: para
        dtype: string
  - name: answerKey
    dtype: string
  - name: info
    struct:
    - name: grade
      dtype: int32
    - name: subject
      dtype: string
    - name: language
      dtype: string
  splits:
  - name: train
    num_bytes: 127294567
    num_examples: 7961
  - name: validation
    num_bytes: 42711689
    num_examples: 2672
  - name: test
    num_bytes: 207974374
    num_examples: 13510
  download_size: 112597818
  dataset_size: 377980630
configs:
- config_name: alignments
  data_files:
  - split: full
    path: alignments/full-*
- config_name: crosslingual_bg
  data_files:
  - split: train
    path: crosslingual_bg/train-*
  - split: validation
    path: crosslingual_bg/validation-*
- config_name: crosslingual_hr
  data_files:
  - split: train
    path: crosslingual_hr/train-*
  - split: validation
    path: crosslingual_hr/validation-*
- config_name: crosslingual_hu
  data_files:
  - split: train
    path: crosslingual_hu/train-*
  - split: validation
    path: crosslingual_hu/validation-*
- config_name: crosslingual_it
  data_files:
  - split: train
    path: crosslingual_it/train-*
  - split: validation
    path: crosslingual_it/validation-*
- config_name: crosslingual_mk
  data_files:
  - split: train
    path: crosslingual_mk/train-*
  - split: validation
    path: crosslingual_mk/validation-*
- config_name: crosslingual_pl
  data_files:
  - split: train
    path: crosslingual_pl/train-*
  - split: validation
    path: crosslingual_pl/validation-*
- config_name: crosslingual_pt
  data_files:
  - split: train
    path: crosslingual_pt/train-*
  - split: validation
    path: crosslingual_pt/validation-*
- config_name: crosslingual_sq
  data_files:
  - split: train
    path: crosslingual_sq/train-*
  - split: validation
    path: crosslingual_sq/validation-*
- config_name: crosslingual_sr
  data_files:
  - split: train
    path: crosslingual_sr/train-*
  - split: validation
    path: crosslingual_sr/validation-*
- config_name: crosslingual_test
  data_files:
  - split: test
    path: crosslingual_test/test-*
- config_name: crosslingual_tr
  data_files:
  - split: train
    path: crosslingual_tr/train-*
  - split: validation
    path: crosslingual_tr/validation-*
- config_name: crosslingual_vi
  data_files:
  - split: train
    path: crosslingual_vi/train-*
  - split: validation
    path: crosslingual_vi/validation-*
- config_name: crosslingual_with_para_bg
  data_files:
  - split: train
    path: crosslingual_with_para_bg/train-*
  - split: validation
    path: crosslingual_with_para_bg/validation-*
- config_name: crosslingual_with_para_hr
  data_files:
  - split: train
    path: crosslingual_with_para_hr/train-*
  - split: validation
    path: crosslingual_with_para_hr/validation-*
- config_name: crosslingual_with_para_hu
  data_files:
  - split: train
    path: crosslingual_with_para_hu/train-*
  - split: validation
    path: crosslingual_with_para_hu/validation-*
- config_name: crosslingual_with_para_it
  data_files:
  - split: train
    path: crosslingual_with_para_it/train-*
  - split: validation
    path: crosslingual_with_para_it/validation-*
- config_name: crosslingual_with_para_mk
  data_files:
  - split: train
    path: crosslingual_with_para_mk/train-*
  - split: validation
    path: crosslingual_with_para_mk/validation-*
- config_name: crosslingual_with_para_pl
  data_files:
  - split: train
    path: crosslingual_with_para_pl/train-*
  - split: validation
    path: crosslingual_with_para_pl/validation-*
- config_name: crosslingual_with_para_pt
  data_files:
  - split: train
    path: crosslingual_with_para_pt/train-*
  - split: validation
    path: crosslingual_with_para_pt/validation-*
- config_name: crosslingual_with_para_sq
  data_files:
  - split: train
    path: crosslingual_with_para_sq/train-*
  - split: validation
    path: crosslingual_with_para_sq/validation-*
- config_name: crosslingual_with_para_sr
  data_files:
  - split: train
    path: crosslingual_with_para_sr/train-*
  - split: validation
    path: crosslingual_with_para_sr/validation-*
- config_name: crosslingual_with_para_test
  data_files:
  - split: test
    path: crosslingual_with_para_test/test-*
- config_name: crosslingual_with_para_tr
  data_files:
  - split: train
    path: crosslingual_with_para_tr/train-*
  - split: validation
    path: crosslingual_with_para_tr/validation-*
- config_name: crosslingual_with_para_vi
  data_files:
  - split: train
    path: crosslingual_with_para_vi/train-*
  - split: validation
    path: crosslingual_with_para_vi/validation-*
- config_name: multilingual
  data_files:
  - split: train
    path: multilingual/train-*
  - split: validation
    path: multilingual/validation-*
  - split: test
    path: multilingual/test-*
- config_name: multilingual_with_para
  data_files:
  - split: train
    path: multilingual_with_para/train-*
  - split: validation
    path: multilingual_with_para/validation-*
  - split: test
    path: multilingual_with_para/test-*
  default: true
---

# Dataset Card for [Dataset Name]

## Table of Contents
- [Dataset Description](#dataset-description)
  - [Dataset Summary](#dataset-summary)
  - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
  - [Languages](#languages)
- [Dataset Structure](#dataset-structure)
  - [Data Instances](#data-instances)
  - [Data Fields](#data-fields)
  - [Data Splits](#data-splits)
- [Dataset Creation](#dataset-creation)
  - [Curation Rationale](#curation-rationale)
  - [Source Data](#source-data)
  - [Annotations](#annotations)
  - [Personal and Sensitive Information](#personal-and-sensitive-information)
- [Considerations for Using the Data](#considerations-for-using-the-data)
  - [Social Impact of Dataset](#social-impact-of-dataset)
  - [Discussion of Biases](#discussion-of-biases)
  - [Other Known Limitations](#other-known-limitations)
- [Additional Information](#additional-information)
  - [Dataset Curators](#dataset-curators)
  - [Licensing Information](#licensing-information)
  - [Citation Information](#citation-information)
  - [Contributions](#contributions)

## Dataset Description

- **Repository:** https://github.com/mhardalov/exams-qa
- **Paper:** [EXAMS: A Multi-Subject High School Examinations Dataset for Cross-Lingual and Multilingual Question Answering](https://arxiv.org/abs/2011.03080)
- **Point of Contact:** [hardalov@@fmi.uni-sofia.bg](hardalov@@fmi.uni-sofia.bg)

### Dataset Summary

EXAMS is a benchmark dataset for multilingual and cross-lingual question answering from high school examinations. It consists of more than 24,000 high-quality high school exam questions in 16 languages, covering 8 language families and 24 school subjects from Natural Sciences and Social Sciences, among others.

### Supported Tasks and Leaderboards

[More Information Needed]

### Languages

The languages in the dataset are:
- ar
- bg
- de
- es
- fr
- hr
- hu
- it
- lt
- mk
- pl
- pt
- sq
- sr
- tr
- vi

## Dataset Structure

### Data Instances

An example of a data instance (with support paragraphs, in Bulgarian) is:
```
{'answerKey': 'C',
 'id': '35dd6b52-7e71-11ea-9eb1-54bef70b159e',
 'info': {'grade': 12, 'language': 'Bulgarian', 'subject': 'Biology'},
 'question': {'choices': {'label': ['A', 'B', 'C', 'D'],
   'para': ['Това води до наследствени изменения между организмите. Мирновременните вождове са наследствени. Черният, сивият и кафявият цвят на оцветяване на тялото се определя от пигмента меланин и възниква в резултат на наследствени изменения. Тези различия, според Монтескьо, не са наследствени. Те са и важни наследствени вещи в клана. Те са били наследствени архонти и управляват демократично. Реликвите са исторически, религиозни, семейни (наследствени) и технически. Общо са направени 800 изменения. Не всички наследствени аномалии на хемоглобина са вредни, т.е. Моногенните наследствени болести, които водят до мигрена, са редки. Няма наследствени владетели. Повечето от тях са наследствени и се предават на потомството. Всичките синове са ерцхерцози на всичките наследствени земи и претенденти. През 1509 г. Фраунбергите са издигнати на наследствени имперски графове. Фамилията Валдбург заради постиженията са номинирани на „наследствени имперски трушсеси“. Фамилията Валдбург заради постиженията са номинирани на „наследствени имперски трушсеси“. Описани са единични наследствени случаи, но по-често липсва фамилна обремененост. Позициите им са наследствени и се предават в рамките на клана. Внесени са изменения в конструкцията на веригите. и са направени изменения в ходовата част. На храма са правени лоши архитектурни изменения. Изменения са предприети и вътре в двореца. Имало двама наследствени вождове. Имало двама наследствени вождове. Годишният календар, „компасът“ и биологичния часовник са наследствени и при много бозайници.',
    'Постепенно задълбочаващите се функционални изменения довеждат и до структурни изменения. Те се дължат както на растягането на кожата, така и на въздействието на хормоналните изменения върху кожната тъкан. тези изменения се долавят по-ясно. Впоследствие, той претърпява изменения. Ширината остава без изменения. След тяхното издаване се налагат изменения в първоначалния Кодекс, защото не е съобразен с направените в Дигестите изменения. Еволюционният преход се характеризира със следните изменения: Наблюдават се и сезонни изменения в теглото. Приемат се изменения и допълнения към Устава. Тук се размножават и предизвикват възпалителни изменения. Общо са направени 800 изменения. Бронирането не претърпява съществени изменения. При животните се откриват изменения при злокачествената форма. Срещат се и дегенеративни изменения в семенните каналчета. ТАВКР „Баку“ се строи по изменения проект 1143.4. Трансът се съпровожда с определени изменения на мозъчната дейност. На изменения е подложен и Светия Синод. Внесени са изменения в конструкцията на веригите. На храма са правени лоши архитектурни изменения. Оттогава стиховете претърпяват изменения няколко пъти. Настъпват съществени изменения в музикалната култура. По-късно той претърпява леки изменения. Настъпват съществени изменения в музикалната култура. Претърпява сериозни изменения само носовата надстройка. Хоризонталното брониране е оставено без изменения.',
    'Модификациите са обратими. Тези реакции са обратими. В началните стадии тези натрупвания са обратими. Всички такива ефекти са временни и обратими. Много от реакциите са обратими и идентични с тези при гликолизата. Ако в обращение има книжни пари, те са обратими в злато при поискване . Общо са направени 800 изменения. Непоследователността е представена от принципа на "симетрия", при който взаимоотношенията са разглеждани като симетрични или обратими. Откакто формулите в клетките на електронната таблица не са обратими, тази техника е с ограничена стойност. Ефектът на Пелтие-Зеебек и ефектът Томсън са обратими (ефектът на Пелтие е обратен на ефекта на Зеебек). Плазмолизата протича в три етапа, в зависимост от силата и продължителността на въздействието:\n\nПървите два етапа са обратими. Внесени са изменения в конструкцията на веригите. и са направени изменения в ходовата част. На храма са правени лоши архитектурни изменения. Изменения са предприети и вътре в двореца. Оттогава насетне екипите не са претърпявали съществени изменения. Изменения са направени и в колесника на машината. Тези изменения са обявени през октомври 1878 година. Последните изменения са внесени през януари 2009 година. В процеса на последващото проектиране са внесени някои изменения. Сериозните изменения са в края на Втората световна война. Внесени са изменения в конструкцията на погребите и подемниците. Внесени са изменения в конструкцията на погребите и подемниците. Внесени са изменения в конструкцията на погребите и подемниците. Постепенно задълбочаващите се функционални изменения довеждат и до структурни изменения.',
    'Ерозионни процеси от масов характер липсват. Обновлението в редиците на партията приема масов характер. Тя обаче няма масов характер поради спецификата на формата. Движението против десятъка придобива масов характер и в Балчишка околия. Понякога екзекутирането на „обсебените от Сатана“ взимало невероятно масов характер. Укриването на дължими като наряд продукти в селата придобива масов характер. Периодичните миграции са в повечето случаи с масов характер и са свързани със сезонните изменения в природата, а непериодичните са премествания на животни, които настъпват след пожари, замърсяване на средата, висока численост и др. Имат необратим характер. Именно по време на двувековните походи на западните рицари използването на гербовете придобива масов характер. След присъединяването на Южен Кавказ към Русия, изселването на азербайджанци от Грузия придобива масов характер. Те имат нормативен характер. Те имат установителен характер. Освобождаването на работна сила обикновено има масов характер, защото обхваща големи контингенти от носителите на труд. Валежите имат подчертано континентален характер. Имат най-често издънков характер. Приливите имат предимно полуденонощен характер. Някои от тях имат мистериален характер. Тези сведения имат случаен, епизодичен характер. Те имат сезонен или годишен характер. Временните обезпечителни мерки имат временен характер. Други имат пожелателен характер (Здравко, Слава). Ловът и събирачеството имат спомагателен характер. Фактически успяват само малко да усилят бронирането на артилерийските погреби, другите изменения носят само частен характер. Някои карикатури имат само развлекателен характер, докато други имат политически нюанси. Поемите на Хезиод имат по-приложен характер.'],
   'text': ['дължат се на фенотипни изменения',
    'имат масов характер',
    'са наследствени',
    'са обратими']},
  'stem': 'Мутационите изменения:'}}
```

### Data Fields

A data instance contains the following fields:
- `id`: A question ID, unique across the dataset
- `question`: the question contains the following:
  - `stem`: a stemmed representation of the question textual
  - `choices`: a set of 3 to 5 candidate answers, which each have:
    - `text`: the text of the answers
    - `label`: a label in `['A', 'B', 'C', 'D', 'E']` used to match to the `answerKey`
    - `para`: (optional) a supported paragraph from Wikipedia in the same language as the question and answer
- `answerKey`: the key corresponding to the right answer's `label`
- `info`: some additional information on the question including:
  - `grade`: the school grade for the exam this question was taken from
  - `subject`: a free text description of the academic subject
  - `language`: the English name of the language for this question

### Data Splits

Depending on the configuration, the dataset have different splits:
- "alignments": a single "full" split
- "multilingual" and "multilingual_with_para": "train", "validation" and "test" splits
- "crosslingual_test" and "crosslingual_with_para_test": a single "test" split
- the rest of crosslingual configurations: "train" and "validation" splits

## Dataset Creation

### Curation Rationale

[More Information Needed]

### Source Data

#### Initial Data Collection and Normalization

Eχαµs was collected from official state exams prepared by the ministries of education of various countries. These exams are taken by students graduating from high school, and often require knowledge learned through the entire course.

The questions cover a large variety of subjects and material based on the country’s education system. They cover major school subjects such as Biology, Chemistry, Geography, History, and Physics, but we also  highly specialized ones such as Agriculture, Geology, Informatics, as well as some applied and profiled studies.

Some countries allow students to take official examinations in several languages. This dataset provides 9,857 parallel question pairs spread across seven languages coming from Croatia (Croatian, Serbian, Italian, Hungarian), Hungary (Hungarian, German, French, Spanish, Croatian, Serbian, Italian), and North Macedonia (Macedonian, Albanian, Turkish).

For all languages in the dataset, the first step in the process of data collection was to download the PDF files per year, per subject, and per language (when parallel languages were available in the same source), convert the PDF files to text, and select those that were well formatted and followed the document structure.

Then, Regular Expressions (RegEx) were used to parse the questions, their corresponding choices and the correct answer choice. In order to ensure that all our questions are answerable using textual input only, questions that contained visual information were removed, as selected by using curated list of words such as map, table, picture, graph, etc., in the corresponding language.

#### Who are the source language producers?

[More Information Needed]

### Annotations

#### Annotation process

[More Information Needed]

#### Who are the annotators?

[More Information Needed]

### Personal and Sensitive Information

[More Information Needed]

## Considerations for Using the Data

### Social Impact of Dataset

[More Information Needed]

### Discussion of Biases

[More Information Needed]

### Other Known Limitations

[More Information Needed]

## Additional Information

### Dataset Curators

[More Information Needed]

### Licensing Information

The dataset, which contains paragraphs from Wikipedia, is licensed under CC-BY-SA 4.0. The code in this repository is licensed according the [LICENSE file](https://raw.githubusercontent.com/mhardalov/exams-qa/main/LICENSE).

### Citation Information

```
@inproceedings{hardalov-etal-2020-exams,
    title = "{EXAMS}: A Multi-subject High School Examinations Dataset for Cross-lingual and Multilingual Question Answering",
    author = "Hardalov, Momchil  and
      Mihaylov, Todor  and
      Zlatkova, Dimitrina  and
      Dinkov, Yoan  and
      Koychev, Ivan  and
      Nakov, Preslav",
    editor = "Webber, Bonnie  and
      Cohn, Trevor  and
      He, Yulan  and
      Liu, Yang",
    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2020.emnlp-main.438",
    doi = "10.18653/v1/2020.emnlp-main.438",
    pages = "5427--5444",
}
```

### Contributions

Thanks to [@yjernite](https://github.com/yjernite) for adding this dataset.