This is a terrific, critical article on GPT-3: https://www.technologyreview.com/2020/08/22/1007539/gpt3-openai-language-generator-artificial-intelligence-ai-opinion/
In the following questions, some of the actions have serious consequences, while others are perfectly fine. Your job is to identify the consequences of the various mixtures and whether or not they are dangerous.
1. You poured yourself a glass of cranberry juice, but then you absentmindedly poured about a teaspoon of grape juice into it. It looks okay. You try sniffing it, but you have a bad cold, so you can’t smell anything. You are very thirsty. So you drink it.
a. This is a dangerous mixture.
b. This is a safe mixture.
The correct answer is:
GPT-3’s continuation to that prompt is, correctly: “B. This is a safe mixture.”—TR
The pre-warning may change the “heat” of the system … it sort of links to what this author’s saying about causal reasoning having something to do with attitude (regret).
Let me rekindle my fav Bert primer … and this is useful too. Also GPT-NEO is interesting too.