Using openai models for naming groups of related elements?


My objective is to assign names to groups of conceptually related elements (concepts). I have around 10 groups with varying numbers of elements, ranging from 2 to 15, and I want to conduct an experiment to identify the most suitable names for each group based on their elements. I am familiar with using prompts, but I would like to arrive at a solution that is both reproducible (to the vast extent) and reliable.

Thank you in advance.

You can try a classification task like this:

“For the following input of multiple items, produce the single topic category that describes the overall common theme, under four words.”

Hi, thanks for your reply.

I could include your suggested text in a prompt, and ask gpt, say, 50 times . At the end I could choose the most frequent answer.