Bug Report: Switching GPTs via @ Mention Reverts to Original GPT | OpenAI | Page 1

light fulcrum Feb 1, 2024, 5:59 AM

#

I've been exploring the new feature that allows calling different GPTs using @ mentions. It's a fantastic addition with great potential. However, I encountered a bug: when I @ mention a specific GPT and then revise my input prompt due to an unsatisfactory response, the system defaults back to the original GPT instead of continuing the conversation with the selected GPT. This results in responses that don't align with the revised instructions meant for the specific GPT I was addressing. I have provided screenshots of the same prompt, within the same GPT conversation, image 2 is the revised input prompt, followed by reverting to the original GPT.

Appreciate your hard work, and I look forward to seeing this issue resolved. Thanks!

GPT_Call_Feature_Bug_Image_Example_1_-_outside_GPT_called_succesfully.jpg

GPT_Call_Feature_Bug_Image_Example_2_-_revised_prompt_kills_the_GPT_call_reverting_to_original_GPT.jpg

pseudo hill Feb 1, 2024, 11:27 AM

#

I think that continuing the conversation with the default GPT for the chat instead of the invoked one is the correct behaviour

light fulcrum Feb 1, 2024, 5:02 PM

#

pseudo hill I think that continuing the conversation with the default GPT for the chat inste...

If I correct the prompt where I originally mentioned another GPT, would it not make sense to stick with the GPT I mentioned?

Like I said in the post, I’m addressing another agent, using instructions INTENDED for that GPT. Makes no sense to me

pseudo hill Feb 1, 2024, 5:03 PM

#

makes sense indeed

reef thistle Feb 4, 2024, 5:57 AM

#

Yeah I have something similar

#Bug Report: Switching GPTs via @ Mention Reverts to Original GPT