Three-year research project investigating metalinguistic capabilities of large language models, focusing on linguistic description, evaluation, multilingual and low-resource language processing, LLM-assisted language pedagogy, and neural language representation methods.