This model is “something else” 🤯

That is blog post is in regards to the first few tests I ran with the recent GPT-4 model. Spoiler alert: It’s impressive!

Chain of thought reasoning

Let’s start with the title image. On this case we ask GPT-4 a really convoluted query that requires multiple chains of reasoning. Not only does it get the reply right, it also provides its reasoning step-by-step!

Logical fallacy

In the following one I attempt to trick GPT-4, and I already thought “Hah, gotcha!” when it non-chalantly added a sneaky footnote into the response 😂

Logic / math — part 1

Any such query ChatGPT got incorrect again and again, but GPT-4 could be very unimpressed by my try and trick it:

Logic / math — part 2

Logic / math — part 3 (with bonus points for the movie query)

This one blew my mind!

AWS exam query — part 1

That’s correct!

AWS exam query — part 2

Also … correct … Wait, am I out of a job now???

Making a sample architecture for a photo-sharing website

Yup, I’m definitely out of a job …

Coding

I’m wondering if I can just let GPT-4 do my job and never tell anyone … 🤪 Can anyone confirm if that is correct, by the best way?

NLP

After all it gets that one right 🙃

Current events

It still has the knowledge cutoff from September 2021, tho. However the geography part is correct, I think (geo was never my strong suit …)?