“Large Language Models (LLMs) are primarily an accumulation of declarative knowledge.”