---
title: FiqhQA
canonical_url: "https://www.modelscope.cn/datasets/MBZUAI/FiqhQA"
md_url: "https://www.modelscope.cn/datasets/MBZUAI/FiqhQA.md"
repository: MBZUAI/FiqhQA
last_updated: 2025-08-13
license: mit
storage_size: "1.1 MB"
downloads: 296
stars: 0
---

# FiqhQA

> FiqhQA - MBZUAI 在 ModelScope 开源的数据集。Dataset Card for Dataset Name

MBZUAI/FiqhQA 是 ModelScope 魔搭社区上的数据集，存储大小 1.1 MB，采用 mit 许可。

- **Repository**: MBZUAI/FiqhQA
- **License**: mit
- **Storage size**: 1.1 MB
- **Downloads**: 296
- **Stars**: 0
- **Last updated**: 2025-08-13

Source: https://www.modelscope.cn/datasets/MBZUAI/FiqhQA

---

# Dataset Card for Dataset Name

<!-- Provide a quick summary of the dataset. -->

Paper: Sacred or Synthetic? Evaluating LLM Reliability and Abstention for Religious Questions

## Dataset Details

### Dataset Description

<!-- Provide a longer summary of what this dataset is. -->

We introduce a novel benchmark FiqhQA focused on the LLM generated Islamic rulings explicitly categorized by
the four major Sunni schools of thought, in both Arabic and English.

- **Curated by:** [Farah Atif, Nursultan Askarbekuly, Kareem Darwish and Monojit Choudhury]
- **Language(s) (NLP):** [ARABIC, ENGLISH]



- **Paper Link :** https://arxiv.org/pdf/2508.08287

## Citation [optional]

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

@misc{atif2025sacredsyntheticevaluatingllm,
      title={Sacred or Synthetic? Evaluating LLM Reliability and Abstention for Religious Questions}, 
      author={Farah Atif and Nursultan Askarbekuly and Kareem Darwish and Monojit Choudhury},
      year={2025},
      eprint={2508.08287},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2508.08287}, 
}


## Dataset Card Contact

[farah.atif@mbzuai.ac.ae]