momahadi/bangladesh-bar-council-exam-dataset

Bangladesh Bar Council Exam QA Dataset: 2022-2023 Bangla-English

The Bangladesh Bar Council Exam QA Dataset is a 400-question bilingual Bangla-English legal question-answering benchmark compiled from the 2022 and 2023 Bangladesh Bar Council examination materials. It is intended for legal NLP evaluation, multiple-choice question answering, retrieval experiments, and Bangladesh law language-model research.

This repository contains the examination benchmark only. The larger fine-tuning dataset is published separately:

[Bangladesh Legal QA Dataset](https://huggingface.co/datasets/momahadi/bangladesh-legal-qa-dataset)

Contents

The English files contain machine-translated versions used in the associated research. Each record includes:

• —question_no
• —question_text
• —options
• —correct_answer
• —correct_answer_text

Loading

from datasets import load_dataset

repo = "momahadi/bangladesh-bar-council-exam-dataset"

bar_2022_bn = load_dataset(repo, "bar_council", split="test_2022_bn")
bar_2022_en = load_dataset(repo, "bar_council", split="test_2022_en")
bar_2023_bn = load_dataset(repo, "bar_council", split="test_2023_bn")
bar_2023_en = load_dataset(repo, "bar_council", split="test_2023_en")

Intended uses

• —Bangladesh Bar Council examination QA evaluation;
• —bilingual Bangla-English legal NLP benchmarking;
• —multiple-choice legal question answering;
• —retrieval and context-use evaluation;
• —reproducible comparison with the associated research paper.

Important limitations

• —The release team compiled and structured the benchmark but did not author the underlying examination questions.
• —The English files contain machine-translated material.
• —Answer keys and translations may contain errors and should be independently checked before high-stakes use.
• —Multiple-choice exam performance is a narrow measure of legal capability.
• —This is a research benchmark, not study advice or legal advice.

Source rights

This repository uses the Hugging Face other license tag. The release team does not claim copyright in the underlying Bangladesh Bar Council examination questions and does not grant rights it does not hold. See `LICENSE.md`.

Separating this benchmark from the CC BY 4.0 fine-tuning dataset prevents the exam material from making the main dataset's license ambiguous.

Contributors and research supervision

The ordered release list is:

1. 1.Moniruzzaman Mahadi - dataset compiler and corresponding maintainer
2. 2.Abrar Mohammed Tanzim Alam - dataset contributor
3. 3.Sayma Siddika Monalisa - dataset contributor
4. 4.Mir Mohammad Asif Abdullah - dataset contributor
5. 5.Mahina Rahman Deya - dataset contributor
6. 6.Swakkhar Shatabda - research supervisor
7. 7.Md Adnan Arefeen - research supervisor

Correspondence: momahadi9664@gmail.com

See `CONTRIBUTORS.md` and `CITATION.bib`.

Citation

This citation is for the compiled and structured benchmark release. It does not represent the release team as authors of the underlying examination questions.

@dataset{mahadi2026bangladeshbarexamqa,
  title     = {Bangladesh Bar Council Exam QA Dataset: 2022--2023 Bangla-English Benchmark},
  author    = {{Moniruzzaman Mahadi} and
               {Abrar Mohammed Tanzim Alam} and
               {Sayma Siddika Monalisa} and
               {Mir Mohammad Asif Abdullah} and
               {Mahina Rahman Deya} and
               {Swakkhar Shatabda} and
               {Md Adnan Arefeen}},
  year      = {2026},
  version   = {1.0.0},
  publisher = {Hugging Face},
  note      = {Compiled and structured benchmark release; underlying examination questions are not authored by the release team},
  url       = {https://huggingface.co/datasets/momahadi/bangladesh-bar-council-exam-dataset}
}