Comment by thwarted
3 hours ago
> "What's your favorite sarcopterygian?"
Am I reading your post correctly, this question is the prompt given to an LLM? What is anyone expecting by asking an LLM what its favorite anything is? This is a conversational prompt, so accuracy and rigor is barely applicable or expected, so downgrading to a lesser model should be acceptable. If you really want to attribute preference to an LLM, consider the downgrade to be a "this conversation is beneath my advanced n-billion parameter training".
I think the intent was just to show how sensitive the classifier is. If it flags prompts that simple, there's no hope for anything biology related at all really.