nassimjp/pashto-kabul-treaty-1921-sft

Dataset Card for Pashto Kabul Treaty 1921 SFT

Table of Contents

• —Dataset Description
• —Dataset Summary
• —Supported Tasks and Leaderboards
• —Languages
• —Dataset Structure
• —Data Instances
• —Data Fields
• —Data Splits
• —Dataset Creation
• —Curation Rationale
• —Source Data
• —Annotations
• —Personal and Sensitive Information
• —Considerations for Using the Data
• —Social Impact of Dataset
• —Discussion of Biases
• —Other Known Limitations
• —Additional Information
• —Dataset Curators
• —Licensing Information
• —Citation Information
• —Contributions

Dataset Description

• —Homepage: Hugging Face Datasets
• —Repository: GitHub
• —Paper: N/A
• —Leaderboard: N/A
• —Point of Contact: nassimjp

Dataset Summary

This dataset contains the complete Pashto translation of the 1921 Treaty between the British and Afghan Governments (also known as the Kabul Treaty), along with 100 question-answer pairs derived from the treaty text. The original treaty was signed at Kabul on November 22, 1921, and ratifications were exchanged on February 6, 1922.

The dataset is designed for Supervised Fine-Tuning (SFT) of Large Language Models (LLMs) in Pashto, particularly for:

• —Historical document understanding
• —Legal and diplomatic text comprehension
• —Question-answering in Pashto
• —Low-resource language model fine-tuning

The treaty consists of:

• —A preamble
• —14 articles covering sovereignty, borders, diplomacy, trade, customs, postal services, and military operations
• —2 schedules (Schedule I: Border adjustments; Schedule II: Legations and consulates)
• —4 appendices (diplomatic letters exchanged between the British Representative and the Afghan Foreign Minister)

Supported Tasks and Leaderboards

• —Text Generation: Fine-tuning models for generating historical, legal, or diplomatic text in Pashto.
• —Question Answering: The 100 Q&A pairs can be used for extractive or generative QA tasks.
• —Machine Translation: The dataset can serve as a parallel corpus (English-Pashto) for translation tasks.

Languages

The dataset is primarily in Pashto (ps), with the original English text provided for reference and alignment.

Dataset Structure

Data Instances

The dataset contains two main components:

1. 1.Full Treaty Text: The complete Pashto translation of the 1921 treaty.
2. 2.Q&A Pairs: 100 questions and answers covering all articles, schedules, and appendices.

Example instance (Q&A):

{
  "question": "د برتانوي او افغان حکومتونو تړون په کوم ځای کې لاسلیک شو؟",
  "answer": "په کابل کې.",
  "article_reference": "Preamble",
  "original_english": "Signed at Kabul, November 22, 1921."
}

Data Fields

• —question: The question in Pashto (string)
• —answer: The answer in Pashto (string)
• —article_reference: The article or section of the treaty referenced (string)
• —original_english: The corresponding original English text (string, optional)
• —category: Topic category (e.g., "border", "diplomacy", "trade", "customs") (string)

Data Splits

The dataset is provided as a single split (train) with:

• —100 Q&A pairs
• —1 full treaty text entry (for generative tasks)

Dataset Creation

Curation Rationale

This dataset was created to address the scarcity of high-quality, structured Pashto text data for training LLMs. Historical documents like the 1921 treaty provide a rich source of formal, legal, and diplomatic language, which is valuable for improving model performance on complex, domain-specific tasks in Pashto.

The 100 Q&A pairs were manually crafted to ensure:

• —Coverage of all major topics in the treaty
• —Linguistic accuracy and naturalness
• —Alignment with the original English text for cross-reference

Source Data

• —Original Document: The 1921 Treaty between the British and Afghan Governments, sourced from British and Afghan archives.
• —Translation: The Pashto translation was done manually by native Pashto speakers with expertise in historical and legal terminology.
• —Q&A Creation: Questions were generated based on the treaty's content to test comprehension at multiple levels (factual, interpretive, and analytical).

Annotations

The dataset was annotated by a team of Pashto linguists and historians. Each Q&A pair was reviewed for:

• —Accuracy of the answer
• —Grammatical correctness in Pashto
• —Relevance to the treaty text
• —Clarity and conciseness

Personal and Sensitive Information

This dataset contains no personal or sensitive information. It is a historical document and does not reference living individuals or confidential matters.

Considerations for Using the Data

Social Impact of Dataset

• —Preservation of Pashto: This dataset contributes to the preservation and digitalization of Pashto historical texts.
• —Cultural Heritage: By making this treaty available in Pashto, the dataset helps Pashto speakers engage with their own history in their native language.
• —Low-Resource Language Support: The dataset supports the development of AI models for Pashto, a language with limited digital resources.

Discussion of Biases

• —Historical Bias: The treaty reflects the power dynamics of the time (British colonialism). Users should be aware that the language and provisions may not reflect modern perspectives on sovereignty and international relations.
• —Translation Bias: The Pashto translation may carry subtle interpretative choices. Users are encouraged to compare with the original English text for critical analysis.
• —Q&A Focus: The questions are designed by the dataset creators and may not cover all possible aspects of the treaty.

Other Known Limitations

• —Small Size: The dataset is relatively small (100 Q&A pairs), making it suitable for fine-tuning but not for large-scale pre-training.
• —Domain-Specific: The content is specific to early 20th-century diplomatic history, which may limit its utility for general-purpose Pashto tasks.
• —Single Translation: Only one Pashto translation is provided; alternative translations may exist.

Additional Information

Dataset Curators

• —Nassim J. P. (nassimjp)
• —Contributors: Afghan historians and Pashto linguists (names withheld for privacy)

Licensing Information

This dataset is released under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0). You are free to:

• —Share — copy and redistribute the material in any medium or format
• —Adapt — remix, transform, and build upon the material

Under the following terms:

• —Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made.
• —NonCommercial — You may not use the material for commercial purposes.
• —ShareAlike — If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.

Citation Information

If you use this dataset in your research or projects, please cite it as:

@misc{pashto-treaty-1921-sft,
  author = {Nassim J. P.},
  title = {Pashto Kabul Treaty 1921 SFT Dataset},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/nassimjp/pashto-kabul-treaty-1921-sft}}
}

Contributions

Contributions to improve the dataset (e.g., additional Q&A pairs, alternative translations, error corrections) are welcome! Please open an issue or pull request on the GitHub repository.

This dataset is part of an ongoing effort to build high-quality Pashto language resources for AI research and cultural preservation.

---