In bug reports, huh?
To get help from OpenAI people in bug reports, best to follow the directions/steps requested here: https://discord.com/channels/974519864045756446/1070242883677982771
#ChatGPT not following detailed instructions
true sequoia
sullen notch
In bug reports, huh?
To get help from OpenAI people in bug reports, best to fol...
i didnt know where to send it since it's so long. I didn't want to make you feel uncomfortable by asking if I can dm it to you so I sent it here. Should I delete it and send it elsewhere?
true sequoia
i didnt know where to send it since it's so long. I didn't want to make you feel...
No idea. You can decide if it's something you wanna bug report 🙂 I linked what this channel is intended for and where the guidelines how to format for here are.
That long information from you could be part of the supporting information you want to use in your bug report, that's fine if so.
sullen notch
No idea. You can decide if it's something you wanna bug report 🙂 I linked wha...
it isn't necessarily a bug, I just wrote it to address what you had initially written about ChatGPT making "assumptions and guess" because of conflicting information. And also I didn't quite understand what you meant about the context window...could you clarify? And also I didn't send the same instructions multiple times, it was supposed to be an automated process since I had specified every little detail in the prompt I had written. But I had to stop it to remind it of the orignal instructions since it kept on going off course.
true sequoia
it isn't necessarily a bug, I just wrote it to address what you had initially wr...
Context window is how many tokens the AI can remember before it forgets the older information. It will try very hard to keep the information you say is more important and let go of less important info, if it can tell.
sullen notch
Context window is how many tokens the AI can remember before it forgets the olde...
Oh? Thanks I didn't know about this at all. Thank you for informing me about this. Now it make's sense how it couldn't remember everything from my resume which was way up at the top and instead made up information on it's own to fill in the gaps instead. Come to think of it whenever I created new chats the first University it created the section for was way better than the ones that followed after it. So if I understand correctly for long tasks such as the one I wanted to do, i should divede the tasks into different tasks and complete them using different chats? In my case that would be creating a new chat for every university?