The LLM couldn’t be enhanced by dynamic training because that’s already what humans do. It’s by design that their “guidelines” are fixed.