Large language models (LLMs) are at the forefront of transforming numerous domains globally. However, their inclusivity and effectiveness remain limited for non-Latin scripts and low-resource languages. This paper tackles the imperative challenge of enhancing the multilingual performance of LLMs, specifically focusing on Generative models. Through systematic investigation and evaluation of diverse languages using popular question-answering (QA) datasets, we present novel techniques that unlock the true potential of LLMs in a polyglot landscape. Our approach encompasses three key strategies that yield remarkable improvements in multilingual proficiency. First, by meticulously optimizing prompts tailored for polyglot LLMs, we unlock their latent capabilities, resulting in substantial performance boosts across languages. Second, we introduce a new hybrid approach that synergizes GPT generation with multilingual embeddings and achieves significant multilingual performance improvement on critical tasks like QA and retrieval. Finally, to further propel the performance of polyglot LLMs, we introduce a novel learning algorithm that dynamically selects the optimal prompt strategy, LLM model, and embeddings per query. This dynamic adaptation maximizes the efficacy of LLMs across languages, outperforming best static and random strategies. Our results show substantial advancements in multilingual understanding and generation across a diverse range of languages.

翻译：大型语言模型（LLMs）正引领全球众多领域的变革。然而，对于非拉丁文字和低资源语言，其包容性和有效性仍然有限。本文致力于解决提升LLMs多语言性能的关键挑战，尤其聚焦于生成模型。通过利用流行的问答（QA）数据集对多种语言进行系统调查与评估，我们提出创新技术，释放LLMs在多语言环境中的真正潜力。我们的方法包含三项关键策略，可在多语言能力方面带来显著提升。首先，通过精心优化针对多语言LLMs的提示词，我们解锁其潜在能力，从而在多种语言上实现性能的大幅提升。其次，我们引入一种新的混合方法，将GPT生成与多语言嵌入协同结合，在问答和检索等关键任务上实现显著的多语言性能改进。最后，为进一步推动多语言LLMs的性能，我们提出一种新颖的学习算法，该算法能够针对每个查询动态选择最优的提示策略、LLM模型和嵌入。这种动态适应性最大化了LLMs跨语言的有效性，优于最佳静态和随机策略。我们的结果表明，在多种不同语言的自然语言理解与生成任务上取得了实质性进展。