是什么驱动多语言语言模型的性能？
What Drives Performance in Multilingual Language Models?
AI总结：
本研究通过在SIB-200数据集上分析6种多语言大语言模型，发现预训练数据规模是已见语言性能的关键影响因素，而文字类型和语系对未见语言更重要，模型规模和架构对核心影响特征无显著改变，为多语言NLP系统开发提供指导。
摘要：
本研究探究了影响多语言大语言模型（MLLMs）在不同语言上性能的因素。我们在SIB-200数据集（一个涵盖204种语言的主题分类数据集）上研究了6种MLLMs，包括掩码语言模型、自回归模型和指令微调大语言模型。我们的分析考虑了三种场景：全部语言、已见语言（存在于模型预训练数据中）和未见语言（未以任何有效方式出现在模型预训练数据中或未被记录）。我们考察了预训练数据规模、通用资源可用性、语系和文字类型等因素对模型性能的影响。决策树分析显示，预训练数据规模是已见语言性能最具影响力的因素。然而有趣的是，文字类型和语系对未见语言至关重要，凸显了跨语言迁移学习的重要性。值得注意的是，模型规模和架构并未显著改变所识别出的最重要特征。我们的发现为当前MLLMs的优势和局限提供了宝贵见解，并有望指导更有效、更公平的多语言NLP系统的开发。
英文摘要：
This study investigates the factors influencing the performance of multilingual large language models (MLLMs) across diverse languages. We study 6 MLLMs, including masked language models, autoregressive models, and instruction-tuned LLMs, on the SIB-200 dataset, a topic classification dataset encompassing 204 languages. Our analysis considers three scenarios: ALL languages, SEEN languages (present in the model's pretraining data), and UNSEEN languages (not present or documented in the model's pretraining data in any meaningful way). We examine the impact of factors such as pretraining data size, general resource availability, language family, and script type on model performance. Decision tree analysis reveals that pretraining data size is the most influential factor for SEEN languages. However, interestingly, script type and language family are crucial for UNSEEN languages, highlighting the importance of cross-lingual transfer learning. Notably, model size and architecture do not significantly alter the most important features identified. Our findings provide valuable insights into the strengths and limitations of current MLLMs and hope to guide the development of more effective and equitable multilingual NLP systems.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
