Beyond LLMs: New Approaches in Natural Language Processing