I came to the forum to get some support, so the first thing I did was search and I found a topic about how to get support, it basically said go to: https://help.openai.com
So I clicked the link and nothing happened except a file got downloaded to my PC named “download” containing a lot of gibberish starting with:
Content-Length: 10464
Age: 1
X-Powered-By: Phusion Passenger(R) 6.0.18
Connection: close
So I tried accessing this site in a different browser, on a different PC and searching from other places to find support and got directed to the same broken link.
I also tried the chat support, it basically just asked automated questions and didn’t provide any feedback.
On the forum there is no support category.
So my questions are:
Is anybody else able to access this support site (you can click the Help Center link under Resources on the forum - gives me the same behaviour)?
Still broken. Returning a file when specifying the main domain.
There was a full day outage of help.openai.com a week or two ago with the same symptom.
You can ask your fellow users here if they know the answer.
You can also do a Google search to get directly into a particular page:
site:help.openai.com gpts faq
If you need to make an inquiry to staff, you can go to platform.openai.com, and press “help” to get the assistant to pop up (only works reliably on Edge browser).
Thanks for all the advice! I will post some questions on the forums, but I recently subscribed to plus and I’ve been having a lot of issues that I felt would require some support from openai. Basically all my prompts end up failing after just a little while and I haven’t been able to get anything useful done, so I would like to know the limits there are on prompt size/complexity and response size/complexity.
For output length, you have a limitation for ChatGPT that has been 1536 tokens, although the AI has been trained to prefer to make you even less than that as output.
It seems like the other concern might be related to comprehension and attention, plus the possibility that you are having an extended chat.
ChatGPT doesn’t have a great memory. You can see all of your conversation, but what isn’t noted is how little of your prior chat is actually sent to the model each time you have another query. So a plan for the best results: you don’t have a long conversation; you form a complete input that alone can be answered by a complete output.
The model also might have a point where your prior input cannot be improved on and consumes all the attention – the AI just gets hung up on repeating back what you wanted rewritten or recoded. So you can take what you need to a new conversation.
I’m getting the message, ‘Something went wrong. If this issue persists please contact us through our help center at help.openai.com.’ and my chat history sidebar won’t load. dsanford0 is the beginning of my email/ username.