Are Dialects Better Prompters? A Case Study on Arabic Subjective Text Classification

This paper investigates the effect of dialectal prompting, variations in prompting script and model fine-tuning on subjective classification in Arabic dialects. To this end, we evaluate the performances of 12 widely used open source LLMs across four tasks and eight benchmark datasets. Our results reveal that specialized fine-tuned models with Arabic and Arabizi scripts dialectal prompts achieve the best results, which provides new findings on the finetuning of LLMs for low-resource languages.