LLMs are contextual meaning they can behave significantly

LLMs are contextual meaning they can behave significantly better or worse depending on the context. This provides opportunity for self-improvement within the same model