Arabic Leaderboards: Introducing Arabic Instruction Following, Updating AraGen, and More

The article introduces the Arabic Instruction Following benchmark, aimed at evaluating models on their ability to follow instructions in Arabic. It also updates the AraGen model, enhancing its capabilities in generating Arabic text. This development is significant for practitioners as it provides standardized evaluation metrics for Arabic language models and improves the performance of generation tasks in Arabic, addressing a gap in multilingual NLP resources.