GPTs and the Brow Detective. Are the reasoning capabilities of… | by Krzysztof Ok. Zdeb | Sep, 2024

Are the reasoning capabilities of OpenAI LLMs adequate to play the basic guessing sport?

GPT taking part in the Brow Detective sport | Picture by DALL·E

Every time I take lengthy highway journeys with my household, we prefer to spend the time taking part in varied spoken phrase video games, particularly those who don’t require an excessive amount of focus from me as the motive force. One in every of our favourite video games is a character-guessing sport. The principles are easy: one individual thinks of a personality — actual or fictional — and the remainder of us attempt to guess who it’s. We are able to solely ask closed questions that may be answered with a easy “sure” or “no”. The one who considered the character should reply with simply these solutions. We take turns asking questions, and the winner is the one who accurately guesses the character first. So it’s fairly just like the “Brow Detective” sport, however with out the Publish-It notes connected to our heads.

Throughout one in all our latest journeys, we had been taking part in this sport, however after a while we began to expire of concepts for characters to guess. That’s once I thought, “Why not let ChatGPT take over this activity?”. I had the ChatGPT app put in on my cellphone, so I launched it and requested ChatGPT to consider a personality for us to guess. I additionally instructed it to reply our guesses concerning the character. And so, we began taking part in with ChatGPT. It got here up with a personality, and we took turns asking questions. The AI…