QIMMA is introduced as a quality-assured Arabic LLM leaderboard that places systematic benchmark validation at its core. It employs a multi-model assessment pipeline combining automated LLM judgment with human review to identify and resolve quality issues in established Arabic benchmarks. The resulting evaluation suite comprises over 52,000 samples, predominantly grounded in native Arabic content, with transparent implementation via LightEval and EvalPlus. Why it matters: This initiative provides a more reliable and reproducible foundation for evaluating Arabic Large Language Models, addressing critical quality concerns in existing benchmarks.
The paper introduces ArabicNumBench, a benchmark for evaluating LLMs on Arabic number reading using both Eastern and Western Arabic numerals. It evaluates 71 models from 10 providers on 210 number reading tasks, using zero-shot, zero-shot CoT, few-shot, and few-shot CoT prompting strategies. The results show substantial performance variation, with few-shot CoT prompting achieving 2.8x higher accuracy than zero-shot approaches. Why it matters: The benchmark establishes baselines for Arabic number comprehension and provides guidance for model selection in production Arabic NLP systems.
MBZUAI researchers release OpenFactCheck, a unified framework to evaluate the factual accuracy of large language models. The framework includes modules for response evaluation, LLM evaluation, and fact-checker evaluation. OpenFactCheck is available as an open-source Python library, a web service, and via GitHub.
The paper introduces NativQA, a language-independent framework for constructing culturally and regionally aligned QA datasets in native languages. Using the framework, the authors created MultiNativQA, a multilingual natural QA dataset consisting of ~64k manually annotated QA pairs in seven languages. The dataset covers queries from native speakers from 9 regions covering 18 topics, and is designed for evaluating and tuning LLMs. Why it matters: The framework and dataset enable the creation of more culturally relevant and effective LLMs for diverse linguistic communities, including those in the Middle East.
MBZUAI researchers have created ArabicMMLU, the first benchmark dataset in Modern Standard Arabic for evaluating language understanding across multiple tasks. The dataset contains over 14,000 multiple-choice questions from school exams across the Arabic-speaking world and addresses the limitations of translated English datasets. It was presented at the 62nd Annual Meeting of the Association for Computational Linguistics in Bangkok. Why it matters: This benchmark enables a more accurate and culturally relevant evaluation of LLMs' capabilities in Arabic, which is crucial for developing AI tailored to the Arab world.
This paper evaluates the performance of GPT-3.5 and GPT-4 on seven Arabic NLP tasks including sentiment analysis, translation, and diacritization. GPT-4 outperforms GPT-3.5 on most tasks. The study provides an analysis of sentiment analysis and introduces a Python interface, Taqyim, for evaluating Arabic NLP tasks. Why it matters: The evaluation of LLMs on Arabic NLP tasks helps to identify strengths and weaknesses, guiding future research and development efforts in the field.