My lecturer’s “solution” to fix LLMs from making mistakes is to make better prompts and provide counterexamples until one mates another. I really want to object that inadequate solution, but I can’t connect the logic with one another…