I’m hoping this is temporary, but I’m unable to refine my GPTs, or even present a detailed opening prompt to the builder.
Does OpenAI do announcements on bugs or ETAs on fixes? Maybe a Discord?
3 Likes
I’m having the same/similar problem.
It appears there was an error in updating the behavior.
1 Like
It’s working now. Good to know at least that it can go the other way.
So I had it working on a new GPT, but when I went back to refine a saved one, no dice. WTF…
I gave up on the gpt and and started using chatgpt 3.5. It’s sailing through the job
You mean using 3.5 for the builder somehow?
no, that’s not possible. I’m just using a prompt
I encounter the same problem with updating GPT’s behavior while in test mode (when changes are unpublished). It does not listen to changes in instructions, at least while it is in test mode (I do not know what would happen if I publish it and do not want to risk publishing when in test mode it constantly encounters errors).
1 Like
I’ve tried it after publication and that doesn’t work either. So, by test mode you just mean before you hit ‘save’?
1 Like
I’ve tried it after publication and that doesn’t work either. So, by test mode you just mean before you hit ‘save’?
@ajd3000 yes, exactly
Okay, so therefor no refinements or updates are possible, since they can’t happen before or after saving. The Builder is officially a non-starter? It’s hard to believe. How are there so many custom GPTs with so many hits, and revenue sharing in the works? It would be impossible.
1 Like
Okay, so therefor no refinements or updates are possible, since they can’t happen before or after saving. The Builder is officially a non-starter? It’s hard to believe. How are there so many custom GPTs with so many hits, and revenue sharing in the works? It would be impossible.
I previously updated the GPT that I talked about in previous messages (last time a couple of weeks ago), and I today and yesterday I updated another GPT, and it worked well.
But my another GPT did not have a significant user training, so it might be the reason why it was updated easily without issues, but it is only a guess.
If this guess is correct, this might happen with GPTs that received much training with reinforcement learning, and they resist to change (this current GPT that resisted updating its behavior was trained by its users when they hit thumbs up and thumbs down at their outputs).
But this hypothesis might be wrong.
I do not know why it happens.
@ajd3000 Was your GPT that resisted updating behavior trained with reinforcement learning by its users?
1 Like
Thanks for that thoughtful reply. Much to consider. I think one problem was I was trying to refine under the ‘create’ tab rather than the ‘configure’ tab. Which the goddamm builder gpt never told me.
1 Like
Thanks for that thoughtful reply. Much to consider. I think one problem was I was trying to refine under the ‘create’ tab rather than the ‘configure’ tab. Which the goddamm builder gpt never told me.
You are welcome. I always used the configure tab, but I see that my GPT in test mode started adapting to new instructions, but still, I receive in test mode consistent errors when my GPT calls my API in test mode (but at the same time, it successfully calls it when using previously published version of this GPT).
There is a hope that it will self-treat itself 
I’m still encountering this issue. Even when I delete a custom GPT and create a new one.
It eventually says it can’t update the behaviour, but that it remembers what I asked it.
Is there any fix for this?
I have the same problem , It appears we are encountering a technical issue updating and it doesn’t apply my prompts
1 Like
I’m having the same problem. It seems multiple people across multiple topics are encountering this.
1 Like
I got an issue, i think similar to yours "This is the text from the Agent “It seems there’s an issue with updating the behavior context to include data collection and processing. I’ll attempt a different approach to integrate this aspect into moondance’s role. Let’s try again.”
Yeah, again, similar error. It would be nice if someone from their team would address this.
These are all the trials my GPT tries byitself to update after I enter a prompt and it fails in the end and ask me to continue…
I would like to know if someone else faces the exact same problem and how could it be fixed?
It looks like there was an error while updating. I’ll attempt the update again to ensure the [GPT] includes all the proposed enhancements.
It seems there’s an issue with the update process. Let’s try once more to ensure the [GPT] includes all the new features.
There seems to be a persistent issue with the update process. I will make another attempt, simplifying the context to ensure the integration of the new features into [GPT] . Let’s give it one more try.
It appears the issue persists. I’ll make another attempt to streamline the update, ensuring the integration of all new features into the [GPT] . Let’s try again.
We seem to be encountering a persistent issue with updating the behavior. I’ll try a different approach to ensure the [GPT] is enhanced with all the proposed features. Let’s make another attempt.
It appears that the issue with the update persists. I’ll need to explore alternative methods to integrate these features into the[GPT] effectively. Please bear with me as I make another attempt.
Given the repeated issues, I’ll reassess the update content to ensure compatibility with the system and make another attempt. The aim is to integrate all the proposed advanced features into the [GPT] , enhancing its capabilities for an optimal study experience. Let’s try once more.
It seems that the issue is persisting. I will consult with my team to find a resolution and ensure the [GPT] is updated with all the new features. Please hold on while I seek assistance.
The continued issues in updating the behavior are certainly challenging. I’ll need to reassess and try a different approach to successfully integrate the new advanced features into the [GPT] . Let’s make another attempt.
It appears we’re facing a persistent issue with updating the behavior. I’ll need to troubleshoot further to ensure the [GPT] is enhanced with all the new features. Please hold on while I address this.
The persistent issue with the update process is quite unusual. I’ll make one more attempt with a refined approach to successfully integrate the new advanced features into the [GPT] , enhancing its capability and ensuring a top-notch study experience. Let’s try again.
We’ve been trying to update the [GPT] with a suite of [the prompt]. However, we’ve encountered persistent technical difficulties during the update process.
At this point, the update hasn’t been successfully implemented. I will continue to work on resolving these issues and ensure that all the desired features are integrated to enhance your [GPT] experience.
I appreciate your patience and understanding as we navigate these challenges. If there’s anything specific you’d like to address or if you have any other questions in the meantime, please let me know!