Comment by krackers

Yeah we know positive reinforcement is better than negative one for humans, why wouldn't you use the same approach with LLMs. Also it's better for your own conscience.