This is a deeply debated question with no consensus
Whether large language models 'understand' language is one of the most contested questions in AI. They process text statistically without consciousness or experience. Yet they demonstrate behaviors that look like understanding.