HalluTruthQA: A Fine-Grained Benchmark for Hallucination Detection, Localization, and Explanation in Arabic Question Answering

Abdessalam Bouchekif, Mohammed-En-Nadhir Zighem, Salah Eddine Bekhouche, Hichem Telli, Somaya Eltanbouly, Shahd Gaben, Heba Sbahi, Samer Rashwani, Mutaz al-Khatib, Emad Soliman Ali Mohamed, Mohammed Ghaly, Abdenour Hadid

Abstract
Large language models (LLMs) can generate fluent Arabic answers, yet factual errors remain difficult to detect, localize, explain, and verify. Existing hallucination benchmarks often provide only response-level labels, offering limited support for identifying the exact erroneous content, explaining why it is incorrect, or recovering the correct factual answer. We introduce HalluTruthQA, a fine-grained benchmark for hallucination evaluation in Arabic question answering. HalluTruthQA contains 2,400 expert-curated examples across four knowledge-intensive domains: Islamic knowledge, history, science, and geography. Each example includes an Arabic question, a model-generated answer, a verified reference answer, a binary hallucination label, and six candidate answers for factual verification. For hallucinated answers, the benchmark additionally provides character-level erroneous spans, human-written explanations, and macro- and micro-level hallucination types. We evaluate four open-source LLMs—ALLaM-7B, Falcon-H1R-7B, Qwen3-32B, and SILMA—in a zero-shot setting across four tasks: hallucination detection, span-level localization, factual verification, and explanation evaluation. The results show that these tasks measure complementary capabilities, with no single model achieving the best performance across all of them. The best observed scores are 0.880 Macro-F1 for hallucination detection, 0.516 F1-Sp for span localization, 0.852 LO-Score for factual verification, and 0.644 for explanation evaluation. Our findings highlight the importance of moving beyond binary hallucination detection toward fine-grained evaluation that also localizes, verifies, categorizes, and explains factual errors. Code, data, prompts, and evaluation scripts are available at https://gitlab.com/nlpresearcher/HalluTruthQA
Anthology ID:
2026.arabicnlp-main.10
Volume:
Proceedings of The Fourth Arabic Natural Language Processing Conference
Month:
October
Year:
2026
Address:
Budapest, Hungary
Editors:
Muhammad Abdul-Mageed, Tamer Elsayed, Houda Bouamor, Samhaa El-Beltagy, Hend Al-Khalifa, Ahmed Abdelali, AbdelRahim Elmadany, Abdellah El Mekki, Kareem Darwish, Sakhar Alkhereyf, Hamzah Luqman, Salam Khalifa, Samar Magdy, Zaid Alyafeai, Injy Hamed, Wajdi Zaghouani, Imed Zitouni
Venue:
ArabicNLP
SIG:
SIGARAB
Publisher:
Association for Computational Linguistics
Note:
Pages:
130–150
Language:
URL:
https://aclanthology.org/2026.arabicnlp-main.10/
DOI:
Bibkey:
Cite (ACL):
Abdessalam Bouchekif, Mohammed-En-Nadhir Zighem, Salah Eddine Bekhouche, Hichem Telli, Somaya Eltanbouly, Shahd Gaben, Heba Sbahi, Samer Rashwani, Mutaz al-Khatib, Emad Soliman Ali Mohamed, Mohammed Ghaly, and Abdenour Hadid. 2026. HalluTruthQA: A Fine-Grained Benchmark for Hallucination Detection, Localization, and Explanation in Arabic Question Answering. In Proceedings of The Fourth Arabic Natural Language Processing Conference, pages 130–150, Budapest, Hungary. Association for Computational Linguistics.
Cite (Informal):
HalluTruthQA: A Fine-Grained Benchmark for Hallucination Detection, Localization, and Explanation in Arabic Question Answering (Bouchekif et al., ArabicNLP 2026)
Copy Citation:
PDF:
https://aclanthology.org/2026.arabicnlp-main.10.pdf
PDF Cite Search Fix data