AraDiCE: Benchmarks for Dialectal and Cultural Capabilities in LLMs
This work addresses the problem of limited dialectal and cultural capabilities in LLMs for Arabic speakers, providing benchmarks and datasets to improve performance, though it is incremental in building on existing translation and evaluation methods.
The authors tackled the underrepresentation of Arabic dialects in Large Language Models by introducing AraDiCE, a benchmark with synthetic datasets for dialectal and cultural evaluation, finding that Arabic-specific models outperform multilingual ones but still face challenges in dialect tasks, with contributions including ≈45K post-edited samples and released resources.
Arabic, with its rich diversity of dialects, remains significantly underrepresented in Large Language Models, particularly in dialectal variations. We address this gap by introducing seven synthetic datasets in dialects alongside Modern Standard Arabic (MSA), created using Machine Translation (MT) combined with human post-editing. We present AraDiCE, a benchmark for Arabic Dialect and Cultural Evaluation. We evaluate LLMs on dialect comprehension and generation, focusing specifically on low-resource Arabic dialects. Additionally, we introduce the first-ever fine-grained benchmark designed to evaluate cultural awareness across the Gulf, Egypt, and Levant regions, providing a novel dimension to LLM evaluation. Our findings demonstrate that while Arabic-specific models like Jais and AceGPT outperform multilingual models on dialectal tasks, significant challenges persist in dialect identification, generation, and translation. This work contributes $\approx$45K post-edited samples, a cultural benchmark, and highlights the importance of tailored training to improve LLM performance in capturing the nuances of diverse Arabic dialects and cultural contexts. We have released the dialectal translation models and benchmarks developed in this study (https://huggingface.co/datasets/QCRI/AraDiCE).