---
license: cc-by-nc-nd-4.0
language: [fa]
tags:
  - multiple-choice
  - evaluation
  - knowledge
  - general-knowledge
  - education
  - persian
pretty_name: PerMMLU Benchmark
---

# PerMMLU

## Dataset Summary

The **PerMMLU Benchmark** is a localized and extended version of the original MMLU (Massive Multitask Language Understanding) benchmark, adapted specifically for Persian-speaking users and domains. This benchmark is designed to comprehensively evaluate both **general** and **specialized knowledge** of language models across a broad range of topics.

This dataset is structured to assess educational and cultural understanding and includes:

- **School-level questions** from **Grade 4 to Grade 12**  
- **University-level questions** covering various academic disciplines  
- **Cultural and local knowledge** specific to Persian-speaking societies, such as:
  - Iranian laws and regulations  
  - Famous historical or contemporary figures  
  - Regional products and traditions

This benchmark is intended to test a model's capabilities in reasoning and factual knowledge in the Persian language, across both **standard curricula** and **region-specific knowledge**.

---

## Supported Task

The benchmark supports the **Multiple Choice Question Answering (MCQA)** task. Each entry includes a question, four possible choices, and one correct answer.

---

## Languages

- `fa` — Persian (Farsi)

---

## License

This dataset is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND 4.0).

You must give appropriate credit, may not use it for commercial purposes, and may not distribute modified versions of the dataset.

For details, see the [license](LICENSE) file or https://creativecommons.org/licenses/by-nc-nd/4.0/