nassimjp/tolanpohena

📘 Tolanpohena — Pashto Social Sciences (ټولنپوهنه) Dataset

A curated Pashto dataset focused on social sciences, sociology, community studies, and human behavior. Designed for Pashto LLM training, SFT, and educational applications.

📑 Overview

Tolanpohena is a high‑quality Pashto dataset containing questions, explanations, definitions, and conceptual discussions related to:

• —ټولنه (Society)
• —ټولنیز جوړښت (Social Structure)
• —کلتور (Culture)
• —ارزښتونه (Values)
• —ټولنیزې اړیکې (Social Relations)
• —ټولنیز بدلون (Social Change)
• —ټولنیز سازمانونه (Institutions)
• —ټولنیز چلند (Behavior)

This dataset is ideal for Pashto LLM fine‑tuning, chat models, educational bots, and academic research.

📦 Dataset Structure

The dataset is stored in JSONL format.

Each entry contains:

{
  "id": "000001",
  "question": "ټولنه څه شی ده؟",
  "answer": "ټولنه د انسانانو هغه منظم ګډ ژوند دی چې ګډ اصول، ارزښتونه، اړیکې او اړتیاوې یې سره نښلوي."
}

📊 Dataset Statistics

🎯 Purpose

This dataset aims to:

• —Strengthen Pashto academic reasoning
• —Improve LLM conceptual understanding
• —Provide clean educational content
• —Support SFT training for Pashto chat models

🛠️ Usage

Load with HuggingFace Datasets:

from datasets import load_dataset

ds = load_dataset("nassimjp/tolanpohena")
print(ds["train"][0])

📚 Content Categories

• —تعریفونه (Definitions)
• —تشریحات (Explanations)
• —مثالونه (Examples)
• —ټولنیز مفاهیم (Social Concepts)
• —پوښتنې/ځوابونه (Q&A)

🔧 Preprocessing

Dataset is:

• —Normalized
• —Deduplicated
• —Clean Pashto script
• —UTF‑8 encoded
• —Ready for training

📜 License

MIT License ---

🤝 Contributions

Pull requests and improvements are welcome.

📬 Contact

For questions or collaboration: Nassim — Spinzar Enterprises Inc.