Using LLM-s for Zero-Shot NER for Morphologically Rich Less-Resourced Languages

Developing Named Entity Recognition (NER) solutions for morphologically rich but low-resource languages like Latvian is a complex task.Most state-of-the-art methods rely on deep learning models like BERT, which require substantial expertise in architectures, methods, and access to extensive computational resources and data.In this study, we explore the potential of using popular large language models (LLMs) in a zero-shot setting without additional training.We evaluate their performance on the publicly available Latvian dataset (Gruzitis, et.al., 2018) using the F1-score and find that their results are comparable to state-of-the-art methods.Moreover, LLMs offer a simpler, more resource-efficient alternative for NER tasks.