Creating Arabic LLM Prompts at Scale
This addresses the lack of Arabic instruction data for LLM training, enabling better performance in Arabic NLP applications, though it is incremental as it builds on existing translation and dataset adaptation techniques.
The paper tackled the problem of generating large-scale Arabic instruction prompts for LLM training by introducing two methods: translating existing English prompt datasets with quality filtering and creating prompts from existing Arabic NLP datasets, resulting in over 67.4 million prompts and enabling a fine-tuned 7B model to outperform a 70B model in Arabic tasks.
The debut of chatGPT and BARD has popularized instruction following text generation using LLMs, where a user can interrogate an LLM using natural language requests and obtain natural language answers that matches their requests. Training LLMs to respond in this manner requires a large number of worked out examples of user requests (aka prompts) with corresponding gold responses. In this paper, we introduce two methods for creating such prompts for Arabic cheaply and quickly. The first methods entails automatically translating existing prompt datasets from English, such as PromptSource and Super-NaturalInstructions, and then using machine translation quality estimation to retain high quality translations only. The second method involves creating natural language prompts on top of existing Arabic NLP datasets. Using these two methods we were able to create more than 67.4 million Arabic prompts that cover a variety of tasks including summarization, headline generation, grammar checking, open/closed question answering, creative writing, etc. We show that fine tuning an open 7 billion parameter large language model, namely base Qwen2 7B, enables it to outperform a state-of-the-art 70 billion parameter instruction tuned model, namely Llama3 70B, in handling Arabic prompts.