If you aren’t getting the results that you want, it might be because you are using the wrong words. A slight change in the words that you are using could change everything.
It can be easy to forget that you are talking to a machine when using tools like ChatGPT. You can use normal human language and the responses seem like normal human language.
BUT - you are talking to a machine. And sometimes it can’t figure out what you mean.
What we SAY and what we MEAN are often different. Sometimes they are really different but another person can easily understand and continue the conversation. If ChatGPT isn’t trained to understand something then it isn’t likely to figure out the actual meaning.
For example, the other day we wanted to have ChatGPT make an image of a bunny for us. The first prompt that we used was similar to “Show me an image of a bunny”. Seemed simple enough.
The response was along the lines of “I don’t have images. Reference a tool like …” and it gave a list of image repositories that we could use.
I initially thought that it was strange because I had used ChatGPT to make images before and I knew that it could do it.
Then we changed the prompt to “Create an image of a bunny”. A few seconds later the image was created.
In this case, it interpreted “show me” as “look at the images that you have stored and retrieve one.” Since it doesn’t store images, it told us “no”. As soon as we asked it to make an image, which it can do, it moved forward normally.
A slight change in words and a completely different result.
If you aren’t getting the results that you want, look back at your prompt and see where your wording could be ambiguous, unclear, or incorrect.