Promoting public literacy in the underlying mechanisms of contemporary large language models