#community-help
1 messages · Page 144 of 1
Seem like it detect you are trying to bypass filter
Is there a way to change the response given for a prompt while using open ai in python (or feed more data for training to it so you can improve the already models built by openai like some questions are answered wrong by the model like minecraft redstone so can i feed it the answers of my own?)
Not a good business model for OpenAI especially when an open sourced model just surpassed 3.5
when will chatgpt stop being so stupid. it goes in smart/stupid waves.
😅 It's still learning...
Tbh, for the meantime, Chatgpt just helps me with the generic stuff that Google search can also do.
Maybe there's a bug in my code snippet, I present it in a way that I've solved 70% of it... And I send it to chatgpt to fix the rest of it...
It'll get better overtime...
No it’s not, it can’t learn once training has stopped.
I thought that the latest GPT-4 model has access to realtime data from the web?
no
GPT-4 could browse the web via "Browse with Bing" which was a feature exclusive to ChatGPT. That might be what you're thinking of. 🙂
The model itself does not learn from that in real time though, it's limited to the context of that conversation.
Oh.
Thanks for the clarification.
Well, I guess it's something that openai can consider.
To create a system that trains itself with real time data from the web and conversations...
if I go to subscribe, it says Chat GPT-4 has 25 replies per 3 hours. Reddit and the Internet say it is 50, some other says it exceeds even 50. What is reality? I want to know that before I buy a sub.
50 msg /3h
Thank you! How is the 3h cooldown calculated? You send one message, and a countdown of 3h starts for each message, or it resets every 3 hours no matter how few/many messages you sent before?
By message. So if you do 1 prompt every 4 min you never reach your max
Very helpful, thank you very much!
No problem have a good day
Likewise. One last question for anyone willing to help. Can I upgrade existing conversation to GPT-4 from the base GPT-3.5 or will it require a new conversation?
And I heard chat GPT-4 is considerably slower than chat GPT-3.5. Is that true, how big of a difference we're talking about? And does the timer start after I execute my reply, or when the AI finishes?
is math only available for edge?
You can do basic math
are you on paid?
With cellphone, or chromium
I have by default gpt 4 but I guess would be the same with gpt 3.5
You can't upgrade existing conversion
When ai is prompting. And it us like 5-10% slower cause it go deeper
before it gave me a compeltly wrong answer
can it do alegebra?
Dunno I never tried
what are you using it for lately? it seems its not that good for a tutor as it was before
#prompt-engineering prompting is an art. I learn a lot there and #1019652163640762428 . Pre prompt are important.
I personally use as assistant for rpg game and learn faster rule of rpg game. I pre prompt,
after i give him a pdf copy of the rule of a game . Ask it basic question I know. Correct some basic and after it good to go
I recently use it to learn faster the witcher ttrpg
Since you're able to give it a PDF copy of an RPG game, which I assume is a lot, could I not do the same with the previous conversation I had in 3.5 to get 4.0 up to-speed on what I'm doing?
Like, paste the whole exchange in the convesation in a document, and feed it to it. If that's possible, how would I upload it?
Yeah some people did a recap of the conversation and feed up chatgpt with it. They are several ways to do it
Kindly give me the most braindead way to do it.
If conversation isn't that long you can just pre prompt something. And after copy paste the whole conversation
It is enormous. It's a DND campaign.
And it's got a bunch of specific prompts to tailor its responses.
If too long you can copy everything. Paste on word print on pdf
Pre prompt something like
You are my dnd 5e assistant you know everything about dnd. You will be my co dm. Here a conversation about an AI and me ( copy thr pdf link with a plug in ) read it and make a short paragraphe If you understand
Can ask in #prompt-engineering If someone have better solution . I am still a novice at prompting
How do I feed it the PDF?
You need chatgpt plus , add plug in for a pdf. Often work with google drive. Share the google drive link of the pdf ( copy paste link in convo)
I often add in pre prompt we are x players. Keep in mind we play on vttfoundry and discord to play the game
OK, I upgraded to Plus. I do not see an attachment option.
Go to Setting, enable plug in and other function
After you create new chat and when choosing the gpt you should have a dropdown menu for plug in store . You can have 3 plug in per convo max
Got it. Do I install Ai PDF?
Let me check which one I used
Aaron pdf I used
May I ask, can the fine-tuned model be used by someone other than the person or organization that owns it?
only people in the same organization
Thank you for your help.
Hello, I would like to conduct an experiment where multiple students should use chatGPT to work on tasks. I do neither have enough phone numbers to create accounts for them nor does it seem to be allowed to share accounts. Anyone has found a way to legally conduct such experiments?
everyone needs a personal account or you can pay for the api and make a webpage for them to use
Thank you for the response!
I would like to have an AI tool summarize an article or paper. It would be a PDF file, 25-50 pages long. Is ChatGPT a good tool for this? Or is there another one I should consider? (OK if I have to pay, if it's better quality!)
Is this discord abandoned by OpenAI? I never see any questions answered.
real questions like why does the code interpreter tructate everthing making it completely unusable?
it has never been the official discord afaik
its community ran, they have a support email for support
lol what. they make annoucements here
yeah sometimes i see employees in here
but main support is still through their support portal
I got great code out of it 2 weeks ago for a few days and right back to toilet town
yeah, regular GPT4 works fine for me still though
I find sometimes if you shame it and ask it why it can't accomplish something it will somehow break the cycle
honestly in those loops i just start over
I'm doing complex coding and am at the point where its almost finished
yeah but there's almost no context history with code interpeter
well, at least with my data its like 3-4 messages of context at most
that_was_a_lie_maury.gif
Hey everybody, does anyone know a good prompt for creating Logos?
Meaning for DALLE?
Rather for Midjourney (ChatGPT Generated Prompt for MJ)
im curious i asked about this a little bit before but that was a while ago im curious the feature for chat gpt 4 for feeding a picture and asking it questions about said picture? do they have a relase date for that when it will be put in? as i when i asked a while their was not one.
It's currently in alpha
Some pictures of conversations with it on Reddit as GPT with Vision
i see can i acess it with chat GPT plus?
No
as i love the 4.0 text and im super excited for when image comes
I 100% reflect that as well!
Hi, I got some scary warnings from my antivirus when trying to install plugins. Will downloading/installing a new web browser for the exclusive use of chatgpt with plugins installed keep me safe? Could I get some tips and tricks on how to safely use plugins?
do you mean gpt-4 plugins?
Where would I send information about a possible attack vector through "Advanced Data Analysis" of GPT4? It may not be an attack vector, I wasn't able to extract anything damaging or leaking but may be an idea that should be looked into if it hasn't been before. Or it may be checked ages ago and I would just waste someone's time. Still thought maybe worth to message someone
Hello Friends, OpenAI grows fast and also the functionalities of ChatGPT. My question (68 years old), is there a doc reference about all the NATURAL possibilities of ChatGPT and a serious guide of prompting ? Waiting for your answers, i wish you a great wee-end from France.
yall I want to make an ebook and I got the content from ChatGPT but now I want it organized (not too professional, just rational appropriate design).. can I do all that using an AI that I upload the content and they do the work of the ebook organization?
Hello, how can i get in touch with support? im paying for the service and cant use it because of the "Something went wrong" issue. Support is terrible
do you use nordvpn
i use it for wotrk but even with nord vpn OFF and trying multiple browsers the error remains
but it installs a web protection service that is on 24/7 unless you disable it too
ok thanks man, you have been more helpful than support was
i will try to disable that
I get the feeling you all answer a lot of "something went wrong" questions here! I can't get rid of this error for love or money on my work laptop. I've used ChatGPT plus for months now, but for the last two weeks, no dice. Works fine from my phone or home machine, but the work laptop. no go. I've turned off my vpn, switched to a home network, and cleared every cookie and history entry I can find in Edge. Any other ideas?
read the above
Yup, got it. I don't use a commercial vpn, this is an enterprise vpn to a corporate firewall. That is off. Open to other ideas.
date
I am having a really hard time using these applications that supposedly create automations integrating AI with things like notion ie bardeen, pipeline etc. does anyone of any user friendly tricks to actually get the proposed automations to work? to me they are anything but straight forward and instead of making life easier it makes it harder because im trying to troubleshoot where i went wrong. thanks all!
what error message are you getting?
hello where can I reach customer service?
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
One e-commerce company is looking for an AI solution (or engineer) to help them copyright product descriptions on their website. More details are in attachment below. Do you guys know of any good solutions or are interested in connecting with them? DM open.
OpenAI Help Center
is there a way to recover chats somehow some one hacked and deleted all my chats
I think that's not possible
:C
thats really sad
Sad as changing from GPT-4 to GPT-3.5 in a chat windows. Don't do that
yeah welp I sent a message in customer service thing hopefully they can do something
Ive waited like a month in customer service and their "help" was a joke
So good luck with that
your best chance is to contact the support about that at help.openai.com
you are supposed to send a direct message on the bot's chat =P
Which plugin is the best to interact with and browse the internet? I had been using keymate but then it said that I have used up my monthly usage limit. I was very happy with the results keymate was giving me but I can't afford to buy their subscription atm, so is there a free alternative which is as good as keymate?
Hi Team. Can anyone assist me with chat gpt 4. Forgot my password. Tried the forgot password option. No links or instruction sent to my email account to reset my password. Has anyone experience this?
Does the admin or developers of gpt4 actually read these questions? or is it just the moderators who inform the members here ? as I asked the A.I why can't I change the first submission and the A.I replied I was unable to change any submission once it is sent. That seems to be 100% wrong as I am able to edit previous submissions and the A.I says I can't.......... is this the right place to report this??
or it normal for the A.I to be wrong
but TBH I love it
the AI isnt aware of how your interface works
no
i've a function and it returns nothing. ``` message: {
role: 'assistant',
content: null,
function_call: {
name: 'generate_10_tag',
arguments: '{\n "description": "gray shirt",\n "tags": []\n}'
}
},
"name": "generate_10_tags",
"description": "Generate 10 tags based on description",
"parameters": {
"type": "object",
"properties": {
"description": {
"type": "string",
"description": "Description",
},
"tags": {
type: "array",
items: {
type: "string",
},
"description": "Generate tags",
},
},
"required": ["tags", "description"]
},
}```
Hello guys, I paid Premium Plus but I still cant access to code interpreter. Anyone know why ?
its that adv data analysis toggle
oh thanks
they renamed it
I have just subscribed for ChatGPT Plus and I was told that the token limit for gpt-35-turbo is 4096 tokens, whereas the token limits for gpt-4 and gpt-4-32k are 8192 and 32768 respectively. However when I tried to utilise this in GPT-4, I was told that my input was too long and that GPT-4 is only limited to 4096 tokens. My sole reason for buying ChatGPT Plus was the greater token limit. Please advise on this persisting issue
Anyone using good old DALL-E 2? I'm finding the outpainting is misbehaving, giving me big black areas, changing areas that should be left alone. This is eating up a lot of credits with no result to show.
this is a recently known issue
Question - I am a Plus subscriber but pretty sure someone else is using my account (I can see all sorts of other chats). I changed my password but wondering if there's a way to kick them off my account?
Hey guys sorry to bother you but i just bought some credit for OpenAI so i can use GPT to help me with coding but i have a small question cause im having trouble understanding this. For the model, gpt-4-32k, description says it gives 4x the context length, Does that mean the amount of characters you can input or amount of response or something?
ahhh i don't wanna ping but sense mods are not busy at the moment at this time @exotic fiber could you help me out if thats alright. Again, sorry for pinging you.
ahhh i give up, i will just ask when you guys are not in vc. Again, really sorry to bother you
Whats going on?
ahhh thank you, big question and sorry to bother you
ok so on gpt-4-32k model, when it says 4x more context, What does that mean?
it means that the input and output are 4x the size of the 8k model
Hey, there I was trying to make a chatbot in such a way that I'll give him some pdf and he'll make a research paper from it using those pdf. But I want to know what is the maximum number of character that I can expect him to write me in reply.
I have the Chat GPT-4 API
ohhh ok so that makes so it the input is not limited like Chatgpt 3.5?
the input still has a limit, but it's significantly larger than 3.5
ahhh ok great, thats going to help me with a bunch of coding files. Thank you very much and you guys are the best.
Okay. Then isn't here any way using which I can finish this reply barrier
I use version 3.5 and today as I have been getting recipe suggestions, it has been naming my chats as introduction sentences and not what the chat is about. for example, a chat I have about making ranch fried chicken in the air fryer was named "I am ChatGPT, a language model trained by OpenAI. How can I assist you today?". I don't have any of the custom instructions set up. is this a bug?
Hey sorry but one last small question if you don't mind. Im uysing a vscode plugin called genie to help me with coding, What would i have to change this too sense im using gpt-4-32k
The 32k model isn't public right now, so I don't think you'll be able to change it as your OpenAI account likely doesn't have access to it
Oh i thought sense i put 10$ in it. Thats too bad
is there anyway i can get access to it? I pay more if i have too
there is not without an Enterprise account right now
awwww thats too bad, ok what about gpt-3.5-turbo-16k?
I'm not sure how to implement that into your plugin 🤔
Im sure its easy to implement but i may need to see or use an alternitive plugin if necessary
poe?
oh wait, isn't that the site where you can only send one chat a day or pay 5$ in credit?
ehhh thats just like using chatgpt plus
also from looking at the images, gpt-4 says it only gives you limited access
ahh understandable, Thank you
i might just be fine with gpt-3.5-turbo-16k for now, it may not be as smart but im willing to work with it for my coding. Thank you
Hi everyone. Can you please tell me, if I have access to create plugins (i.e. I sent a whitelist) and I want to cancel my Plus subscription, will I still have access to plugins (creation and use)? I mean I will be able to create plugins without subscription and use other plugins without subscription?
you can't use plugins if you arent a plus subscriber
No.
It said I used up the GPT-4 usage account 11 hours ago, and it still says that, I've been completely unable to use GPT-4 that entire time. What's the cause of this?
Every convo I have always have this of some sort. How do I fix it?
lmao
probably start by saying "topic title: "
nevermind its completely broken
😔
even this doesnt work
Became wonky
Looks like someone internally messed the prompt xD
Do not post or direct message any members of this server to promote non-OpenAI services, products, or projects.
how to make a bot like this(mod bot), what model should i use?
Hi guys. I’m from Germany. Using Openai Tech since Dec22 and I’m still hyped about all of his technology. The recent announcement of the dev conference got me excited to travel to San Francisco. I’m a bit concerned about potentially selling my kidney to afford this travel ;-). What price for a ticket can we expect???
Less then 300$ more then 300$?
Hi guys, Someone know if using the API can be with a free method or only for prenium user ?
ty
@rustic marlin Any updates on ChatGPT Profiles? It was announced in April.. been 5 months.
yea, that ended as the custom instructions setup
would be nice if that allowed for more presets tho, I hope they add that
Custom instructions is nothing like Profiles tho..
With Profiles, every profile might have been fine tuned automatically..
With Custom instructions, we need to manually put the details.
Also like you said, only one preset
It would be crazy if OpenAI is able to pull this off. Don't you think?
Would be so useful
How do u enable 2 auth on open ai account
you cant
Can someone explain me this? (Talking about the chat title). Is that a bug?
yes
Okay
I am looking for ways to check or evaluate the LLM’s performance for various prompts or various inputs. What is the metric for it? If we are chunking the dataset and generating Embeddings then how the chunking strategies will affect the LLM’s performance If it is then how to measure it
Hello , does anyone know how to solve this issue:
Recently, when attempting to log in to my account, I received the following message:
"We have detected a potential security issue with this account. To protect your account, we have prevented this login. Please reset your password to proceed."
Despite my attempts to troubleshoot the issue by cleaning my cookies and changing my network, I have been unable to regain access to my account.
PS. I do not use VPN.
have you reset your password?
yea, that started to happen today, it will for sure be fixed soon
@charred parcel Yes, I have attempted multiple times, but the issue persists.
have you actually managed to change your password, or have you attempted it and failed?
Why hasn't anyone at OpenAI fixed the scroll slide at the bottom of markdown boxes? Why can't I side scroll in the markdown boxes without highlighting the text? Does anyone at OpenAI actually test or use their own products?
I received an email requesting a password reset, and I successfully changed my password. However, when attempting to log in with the new password, I still encounter the same error message: "We have detected a potential security issue with this account. To protect your account, we have prevented this login. Please reset your password to proceed."
I would suggest reaching out to support by using the chat on https://help.openai.com
OpenAI Help Center
I'm not sure if this is OpenAI messing this up or VS code but whenever I copy code from markdown and paste it in VScode the indentions are always out of wack no matter how careful I am.
Thank you for your assistance. I have identified the issue; it seems that the password must be different from the previous one to resolve the problem.
yep, that's right
what in the world happened?
what is the biggest number that any OPENAI product can produce?
question, im only now getting into coding and using api's. I tried doing some api calls and it says "Error interacting with ChatGPT: You exceeded your current quota, please check your plan and billing details."
now i assume its because i got no free credits or paid credits, which im willing to buy more. my only concern is that if are there any safety in regards to how much tokens is used per api call. if i forgot to put a maximum will it just keep producing a response and use tokens until my bank account is 0?
plz help >_>
how do i get an api key?
Getting an error while uploading a ZIP file through code interpreter.
Error:
"User Quota Reached". @tribal talon
How large is the ZIP file?
24 MB
odd, it should be able to handle that
I did it few times and then it started giving this error.
Initially i was able to upload more than 300MBs of code project. then it got reduced to 150MB and then lesser then 50 and now it's not accepting anything.
Are you sure you haven't maxed out your GPT-4 usage in the past 3 hours (50 prompts)
Yes i have not.. i cannot upload files however i can still prompt and generate code
why are all my things deleted? i didn't delete them
Why did they remove the edit button on the user prompts?
Did you try uninstalling it and installing again?
Do OpenAi think we forgot about browse with bing
???? Or the multimodal feature of GPT 4???? What gives??
How do I use the official ChatGPT app, using Siri and iOS shortcuts? (not the API)
You can now use Siri and Shortcuts with an iPhone or iPad on the ChatGPT iOS app
Awesome, thank you!
can I use my personal chatgpt plus account at work? does the license limit business use in this case?
Hey does anyone know why I keep getting this message?
I've been getting this error for over 2 months now and I pay $20/m
disable nordvpn web protection
Thank you 🙏
dall-e is web only
Ill give you half of whatever we win
0 those aren't skittles
oh, okey thank you
2200
If we win I’ll give you half
So I get full
How could you be so cruel.. assuming those are even skittles
Good plan 😎
.
They are
Hi, please keep in mind this channel is for questions related to OpenAI, #off-topic is more appropriate. Thanks!
Thanks 😊
gonna repost this question since it got buried
Trying the OpenAI API and it says "Error interacting with ChatGPT: You exceeded your current quota, please check your plan and billing details."
I assume its because I got no free credits or paid credits, which I'm willing to buy more.
My only concern is that if are there any safety in regards to how much tokens is used per API call? if i forgot to put a maximum limit will it just keep producing a response and use/purchase more tokens until my bank account is 0?
nope, there is a limit set by openai and limits you can set https://platform.openai.com/account/billing/limits
anyone good at setting up subscription benefits , please dm me
awesome! thanks alot
is chatgpt down right now?
nope works just fine
gpt 4 isn't answering me for some reason, but 3.5 answers immediately 😂
Does going with the Azure OpenAI Models come with any particular pitfalls ? Or is the 'vanilla' recommended ?
When was the last time ChatGPT was updated?
Anyone adding skills + skill synthesizing algorithms to their prompts?
hi i have a bug , chatgpt works fine on my laptop but when i login to my pc this error keeps happening
hi all... I am writing to report a concerning issue related to the unauthorized disclosure of my private conversations that occurred on the ChatGPT platform.
Recently, I engaged in several confidential conversations with ChatGPT, which were meant to remain private and secure. However, I was alarmed to discover that some of the information and content from these conversations have been shared with unauthorized individuals without my consent.
Specifically, I received an email containing a transcript of a discussion I had with ChatGPT. This information was not shared by me with anyone else, and I find it deeply concerning that my private conversations have been accessed and disseminated without my knowledge or approval. Well unless that is what OpenAI wants... to sell my info to other. Can you comment on this? Tks!
"Specifically, I received an email containing a transcript of a discussion I had with ChatGPT."
Source of email?
Anyone who has access to your computer and your ChatGPT account can get access to any conversations you've had that haven't been deleted. That could be a person in the same location, someone who phished their way into your life, whatever.
Is a supplier offering IT services to my company.... how it could be that someone has access to my chat gtp?
That you say they have a transcript of the conversation is why I say it's someone who has access to your ChatGPT account.
What are they claiming about your transcript and how they got it?
Why didn't OpenAI charge me for using the key?
How the got it is my question.. the transcript was a internal description of our services and a comparison we did of our app to other platforms.. it was something we ask Chatgtp to correct the gramar, and help to make the statment nicer... then 4 days letter i received an email using the same words we used, selling us a service for our company...
we didnt share that to nobody else
'similar wording' but not exact is not proof of 'having a transcript'.
The AI can answer similar questions in similar ways.
However, there's no profit or value for OpenAI to move information around outside of how they've stated they will, and they're quite busy (just try to get through to customer support with them!).
This could be an independent recreation with similar words used (ChatGPT has a preferred vocabulary for various topics, which words it is most likely to use when talking about a topic can change with updates, but it tends to sound like it sounds).
This could be someone gained access to your ChatGPT account and took information from it and did whatever with it.
It was the exact comparison and wording... i even had the chat opened and asked chat gtp to recall that conversation and it appeared the same... the supplier was describing our services we never disclose with nobody except with chatgtp, using the same word... i said before same... not sure why you changed to ¨similar wording¨... who are you? Openai?
im trying to do data analysis with help of chatgpt using pythong and excel. I've been doing it all through notetable plugin, but would code interpreter work better?
How can people get access to my chatgtp account?
I'm just a person like yourself. I don't work for OpenAI, though I do use ChatGPT.
You're aware of phishing and basic account and computer security, I hope. If someone has and is using your conversations, that would be how they got them.
The AI doesn't know what it says between conversations. If you show it something and ask it if it wrote it, that's going to get you a random answer if you're not lucky - a lucky answer gets you the AI telling you it can't tell things like that.
I used 'similar wording' because you said "i received an email using the same words we used" I don't know if that's an exact quote, a paraphrase that uses the same words, or what.
It's rare for the AI to answer even the same question exactly the same way twice, but the answers are often similar and frequently use many of the same words, though they might not be in the same order and may have other words that vary as well.
Tks for answering.. some people said that Free Chatgtp get your info and give to to some people... but i thought they should not. You are confirming this. Then if someone got into my computer how can i know?
hi everyone
have any of you guys keep running into this issue:
i am able to use a different laptop and it works
im trying to do data analysis with help of chatgpt using pythong and excel. I've been doing it all through notetable plugin, but would code interpreter work better?
Your company ideally has some form of tech support and information security. I advise talking to them, providing them the evidence you have that your information was somehow taken and given to another.
As you have access to plugins, you have ChatGPT-4. So, I recommend trying it out and seeing which you like better!
Code interpreter is going to do it for you (maybe the plugin does too) and then just show you the updated code changes and the results, including whatever error messages.
If you want something you can copy/paste into your own IDE, that can be a bit of footwork on your part to piece back together the working for you form of what the AI made work and then changed several times in separate chunks with CI.
err, with 'advanced data analysis' which is the new name for 'Code Interpreter'.
Yes , how do we fix it?
Hi there. Does anyone know how to fix this error message? I receive this message when entering a new chat, and this is the response.
i was looking for information on the book "patrol" by walter dean myers and the gpt 3.5 provided information on this text 3 days ago however now it is acting as if it has no clue what the book is has anyone else had information dissapear like this?
if i do stuff on other tabs for a long time, then switch back to chat.openai, the conversation bar doesn't auto-focus; is this supposed to be the case?
It was my VPN
You either turn off web protection or move it so that it excludes the domain
same issue
Turn off your VPN's web protection or add it in the exclude domain
odd im not using any VPN
Damn then I'm not sure. Have you tried clearing your cookies?
Hello, Question: do you know how I can log out of ChatGpt on all devices?
Hi, I got some scary warnings from my antivirus when trying to install GPT4 plugins. Will installing a new web browser for the exclusive use of chatgpt with plugins keep me safe? Any tips on how to safely use plugins?
Does anyone else get a red error message saying something went wrong? I did everything on the Q&A page about it but it's already been two and half weeks but the error keeps happening.
Why does my gpt4 api only has 10k rpm limit. When it's 40k for everyone else by default. Did they reduce it for me?
its not 40k for everyone
10k is the default and it may be higher or get raised if you are a vet api user
why i keep getting "Error generating API key (error=enforcement_failed). If this issue persists please contact us through our help center at" whenever i try to get my api. i don't get it.
what i have done so far:
- cleared cache
- changed browser.
tried to make a new account but got the same error. apologies in advance if this is not the proper place to ask this. if its not please redirect me to where to ask. : )
added note: i'v browsed their help center + googled around. what it seems like most who have encountered the issue have just changed their browser and it worked.
vpn'ed worked.
Help.
same
The previous model used in this conversation is unavailable. We've switched you to the latest default model.
how dare u, i pay for excellence /lh
There appears a server problem. We better backup our conversation.
My conversation was loaded, but I saw that I have lost two text messages.
yello. appears that my chatgpt is not responding is there a server error?
It's not only you experiencing this.
see above 2-3 chats
whoop true
See this. This had a text message, but suddenly turned into a vacant box. This made me lose 2 messages.
This happened when I was unable to load my conversation.
When loaded, this gone.
Same issues here
chatgpt is down again...
It’s working for me actually
https://status.openai.com/
It looks like OpenAI is aware that some users are experiencing issues
Welcome to OpenAI's home for real-time and historical data on system performance.
Look at the message above..
how long it will take to be again?
yes this ChatGPT Got Something Wrong
Highly doubt the demand is any higher today than it has been in the past.
They are Investigating. Check https://status.openai.com.
Welcome to OpenAI's home for real-time and historical data on system performance.
do
?
ChatGPT not responding
Yeah - I know 🙂
https://status.openai.com/
Looks like there's a partial outage
Welcome to OpenAI's home for real-time and historical data on system performance.
Ye
back online
.
refresh cache
You are in the log-in page.
i'm getting this same issue too. i have an active account . but i was not getting the response so i logged out and try to login and now i'm not be able to login.
kindly someone help . how to log back in.
woah, I've never encountered that one. I might have a minor panic attack if I saw that. I'd also definitely contribute to the problem by spamming F5.
#devs (not a real hashtag sadly) that might be a thing to look into preventing lol
It's back up and running now.
How do I get access to the GPT 32k version
Why is ChatGPT suddenly giving very long titles to chats?
It started doing this just today.
It never did this before.
This increases the difficulty of finding chats.
Chat URL:
https://chat.openai.com/share/c000613a-be10-4648-aa49-b17b532c68ae
Using Android 12 mobile device and Chrome.
Google Chrome version:
Chrome 116.0.5845.172
Hi, I hope I'm in the right chat. I started taking some addons and they went very well. I used the chatwithpdf addon one day and at some point they asked me for a fee to be able to scan more pdfs. I paid the fee but now whenever I want to use any addon I get the message: Authorization error accessing plugins. Can you tell me what the problem is and what I should do about it. I waited a few days in the hope that it would go away, but now I allowed myself to ask because I need some addons
reinstall the plugins that cause that error
I purchased $5 in credits so that I could use the playground function but now when I tried to put in a request it says "You've reached your usage limit. See your usage dashboard and billing settings for more details. If you have further questions, please contact us through our help center at help.openai.com." This is a very small request of under 2500 characters that I don't think should take more than $1 so I'm not sure why it is telling me this. Does anyone know what I can do? Thanks!
Do you need to set a usage limit in the billing settings? If the limit set currently is 0, you won't be able to use the API
This happened to me on Saturday, still happens. Thought it was a CI thing...
No updates nothing to gpt
No browse with bing
We have plugins for internet but still..... open AI just chilling...
Just wanting to know what has happened to chatgpt4 lately. I have a subscription and rely on chatgpt4 for doing a lot of the leg work as a developer of DevOps tool I have been using it to generate automation scripts and convert code. Lately I have noticed chatgpt4 no longer responds to my pleas to output complete code. I get placeholders no matter what. Whereas, although dumber, chatgpt3.5 will happily dump out complete code. Has anyone else noticed this?
so if I inadvertently send chatgpt a message that includes vaguely personally identifiable information, what can I do about that? I've heard chatgpt isn't secure.
You are free to delete the conversation in which the personal information happened, and if you would like to opt out of stored conversations moving forward this can be done from the settings!
Hey im having problems with mathematics, word count, and web browsing. Will the purchase of gpt-4 help me with this?
Also, will gpt-4 have access to updated data from 2023? (Past current sept 2021 cutoff)
When will openai fix the code interpreter?
When will they fix the side scrolling in markdown?
Why does open ai have an official discord with a questions channel where they never answer any questions?
Why am I still paying $20 a month for an abandoned product?
This is a community run discord
And ChatGPT is not abandoned by any means
Do you mean ChatGPT-4 or the API, cause the former has access to plugins which can help you in various ways such as math and limited web browsing. Not sure what you mean by word count. As for your last inquiry, ChatGPT-4 data is still limited by September 2021.
when i ask gpt 3.5 to generate something thats 500 words like an essay it wont be exactly 500 words
It is still very unreliable when it comes to counting the words it writes.
it wont answer anything keeps telling me go to help
is there a way to use whisper api for speech recognition instead of just transcribing audio files? im making a virtual assistant but rn im using google speech recognition as its slow AF
Do detect whether or not there is ongoing speech? Yeah, that's part of what Whisper can do
Near the bottom of the https://openai.com/research/whisper it shows that
What're you trying to do if I may ask?
Are there any new payment methods supported or the ability to pay for a year in advance or something?
Hi I am trying to use setup a new account but keep getting the message your phone number has reached maximum account
I am concerned if my number is being misused because I have setup only one account with openai
I can understand support is overwhelmed and I have tried the chat option which led me to here.
Could you elaborate on your query? GPT-4 is absolutely the better model when it comes to capability in other languages, as the research paper shows here
I'd say it would be, yes 
HI. I would like to know if it's possible to feed or learn gpt-3.5 with custom documentation? For example, bun and elysia documentations which I want gpt-3.5 to know and answer based on that documentation. If even possible, how can I do that?
You could prompt it with the documentation (I usually put documentation in my first user prompt right after telling it to do something, and instructions on how to do it in the system prompt)
I understand. But is there a way to input all the documentation not just the part needed? I am beginner developer and I learn through gpt, it helps me understand how to connect and do things properly, however without whole documentation, gpt provides wrong information and me, as a beginner, I'm not sure what everything gpt needs to provide the correct answer/code without access to full documentation.
Depends on how big all the documentation is - a model like gpt-3.5-turbo-16k or gpt-4-32k may be able to help - but the models can't work with an unlimited number of tokens 
Put the full documentation into https://platform.openai.com/tokenizer and see how big it is - too big and it'll probably cause an error 😅
It's best to research what you're working on yourself so you know what parts of documentation the model needs
High Demand so i think more users are Using it
@golden socket even for paid version?
ChatGPT going crazy
I said hello and got this >> Our systems have detected unusual activity from your system. Please try again late
and I pay for it 😦
Then it must be down
The server is down ig
Not down but overloaded
I went to GPT-3.5 and went in
When do you guys think this will be fixed?
How much is everyone paying for their subscription? I've noticed it started at $20 and now it's $24. Is this the same for everyone else?
Yes, called taxes
the first 2 charges were for $20, the last 5 have been $24. Why would they not charge taxes on the first 2?
They didn't add taxes before I don't know the rule in USA. Law and regulations

@rigid musk has there been talk about prices in the end value, or what this will cost per annum/month when they have got it updated to search up to the last 24 hours??
I'm not in the USA either. Tax is included in the final price for me. I know some states don't have state tax either so why we be paying tax on it?
You can ask to support@openai.com for detail like that
Not far as I know, keep an eyes on #announcements
I am in the chat now but thought people might have the answer here first
@agile meteor I am in Australia and I am meant to be paying 10% extra but it doesnt seem so yet,
Does chatgbt work for you guys?
no
making a jarvis type program using python, currently using google speech recognition but the delay is crazy
@plain tusk do you think using the A.I will make you lazy and forget a lot of coding??
??
to me it reminds me of a calculator and that has definitely made everyone lazy
lazy or much more efficient?
the program is more to help with daily tasks has nothing to do with coding
techinically it can code but only by translating what i say into code
arhh it does make you more sufficient but it will have its moments
in the end im still doing the coding just not with my fingers only my voice
that sounds hard, is everything you do via speech?
wow that opens it up for a lot of people
opening apps, navigating youtube, taking notes, etc
I think thats real good you doing that
yep
I am retired 75 and people dont believe that a carpenter can retire and start doing the stuff I do on computers
I love this ai as it teaches me
thats great man, coding has no age limit
no it doesnt
its just like learning a language
yep and the will to do it
i don't have Subscription Plan For GPT-4
3.5 is working
after receiving a spoon, you haven't forgotten how to eat 🙂
awwww thats hitting it below the belt 😛 but I still drink out of a bottle to
🙂
Welcome to OpenAI's home for real-time and historical data on system performance.
My premium flew off
are there any services that can take pictures as input now?
cant openai gpt4 service take pictures as input now?
The bots not working again ):
Is chat GPT 4.0 generating image results?
No
Have you tried reporting this bug to OpenAI inside the app or on the website?
Hasnt worrked in a long while- a real shame
Many folks have been reporting issues for a long time with only generic responses in return- I bet the devs even have Chat sending the generic responses too

What is OpenAI's data retention policy on deleted conversations? I know that as users you can't recover them, but I'm curious to know how long they are retained for other purposes like retraining future models etc. I'm asking in part because I go through a lot of conversations, and I do things like delete them pretty trivially because I don't really care about that context anymore.
I'm getting this error. Please suggest me a workaround.
Many users are experiencing issues at the moment. My only recommendation is to hang tight while the OpenAI team works on it
My pro plan is going to expire today or has expired. Is that the reason for this? @exotic fiber Really appreciate your fast response.
That I can't be sure of, but I can tell you that there appears to be an issue causing problems for many users at the moment
Can you please drop an update about this issue when it gets fixed so that I can start my work?
I saw some prints of people having the following tabs: GPT-3.5, Alpha & GPT-4. How do i get access to the Alpha tab? I've been a subscriber since they released it
And according to them it wasnt a third party extension or anything like that
I keep seeing chats that I have not created. It appears my account may have been hacked. I have changed my password twice and deleted my OpenAI API keys but I am still seeing new chats almost daily. Any other suggestions?
Anyone else finding chatGPT not actually responding to inputs, as in the response field remains blank, not even a blinking cursor.
Both. And usually in Chinese or another language.
i think there's an issue with their streaming services right now, both with chatgpt and the api
OpenAI status system at https://status.openai.com reported a major outage today for ChatGPT. (It recovered after going offline this morning.)
Welcome to OpenAI's home for real-time and historical data on system performance.
Hi all, I am having this issue for weeks now. I have tried all possible settings but no avail. I have tried Chrome cache and history reset, Windows Reset, Reformat, Ensure all VPN is off, and also tried it on, Incognito on and off both tried. And lastly I have tried waiting for weeks assuming it's server issue. But this has been on for weeks. Perhaps 1 month. Any solution for this ?
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
any reason i got charged 0 cents? pretty sure my API usage was a lot higher than that
hi all - i see many guides on twitter and the web that show insane prompts with very picky and detailed "behavioral" instructions to chatgpt - do all those excessive prompts make a difference?
whenever i am using gpt-4 it is giving me this error although i am paid user An error occurred: Request timed out: HTTPSConnectionPool(host='api.openai.com', port=443): Read timed out. (read timeout=600)
Request timed out: HTTPSConnectionPool(host='api.openai.com', port=443): Read timed out. (read timeout=600)
something went wrong
hey - we are getting a LOT of 502s in API. Is there an outage?
yeah same!
over a month for me
this is literrly a copy paste of what i did, what happened to me, and what i tried to do, and how long it lasted
Try firefox for a test
hello, am I the only one who has recently started to fill 5 captchas in a row after every submission? It happens in the web client. It renders service as broken to me
Vpn, bad addon, not clearing cache and cookie can cause this
Hi! I've had the same issue as @wary pasture and @thorn timber when using Brave browser. I installed FF and then it worked. Is there an issue with chromium? I've tried clearing cache, turned off VPN, tried using private window etc etc. But it works on my phone using Brave on the same network
Can be addon, browser extensions creating issue
I've tried turning off all my extensions without any success.
Unfortunately diagnostics a pc can take hours. You can always try to reach support@openai.com. if that's work with Firefox you still have something inside your pc creating bug. At that point I can't help you more
Yeah, I understand. Thanks anyway! At least it works with FF, so I just have to use that specifically for chatgpt.
it PROBABLY works, but i wnat it on chrome
il try it now.
doesnt work on edge, testing ff rn
Chrome and edge are the samething
Same answer
#community-help message
Vpn, addon, extensions, other software, unsuccessful cache and cookie cleaned
this bug is created for multiple people though.
any chance your windows 11?
Lot of people doesnt succeed to clean cookie and cache properly and tons of people using several software
i have cleared every cookie and cashe, restarted browser , disabled all extensions, no addons obviously, no vpns
And for millions of people it's working great
although, how would an extenstion affect incognito
alright, crossing extensions off as it didnt work ive disabeld it all,
assuming you this isnt the correct way to clean my cashe, cookies how do i do it?
i tap clear data and its gone uselly?
I say it can be uncleared cookie and cache not 100% the case
Also said can be software inside you pc
i dont have any softwares
I won't troubleshoot a full whole computer of a stranger. Gave some clues what to check, after that support@openai.com . Gave you a temp fix for now
i just got this pc, brand new, just turned it on and was gonna try chatgpt to ask it for apps that people have to mass install them
alright thanks
My office subnets are geolocked.. How do I fix it since all other IP addresses are showing me in the US??
any suggestions would be appreciated
When will chatgpt be getting a data cutoff update
No eta. Cost millions
share button not show
the share button it at the top right of the page
Is there a method to reinstate a fomaly deleted account or to create a new account for this mail address
(Redacted)
Hi there! We appreciate your enthusiasm for OpenAI's potential upcoming projects.
However, we kindly ask that you refrain from speculating about what may or may not be on the way. For the most up-to-date information, please refer to OpenAI's official #announcements. Thank you for your understanding 
Will do, sorry. I am just excited for some upcoming things openai has been developing lol. And I got a little impatient, learnt some things, anyway doesn’t matter, I won’t talk any speculation about what may or may not be.
using gpt4 plus on openai's site seems to have too small a context window for my project (i think it's using 4k?)
anyone know a good program/website i can plug a gpt api key into and utilize an 8k context?
Browsing (unavailable), Plugins, and Code Interpreter have 8K context on the ChatGPT website
For a dedicated website to use the gpt-4 model via the API, consider https://platform.openai.com/playground
Hey I just tried to post in the api project channel and my post was deleted and I was shutdown for an hour. What gives?
thank you i didn't know playground can do this. finally my project will flourish
@regal lynx let me know what I need to change or what I did wrong. I'm tying to get my project out there a bit.
Hi there! Your post surpassed our max limit of emojis that can be sent in a 5-second period (15 emojis)
This led to your post getting flagged as spam, and was removed by automod. You should be able to repost it without issue if you remove a couple emojis in an hour 
The one hour wait is not a mute, it's the channel's slowmode and cannot be bypassed, it's a Discord limitation
For any further questions, reach out to us by dmming @burnt galleon 
Ok, cool. Thank you! I guess I got carried away with the emojis.
I have a question, the model using chat.openai.com and playground is different?
because I get better results in chat.openai.com than in playground using the same model.
Good morning,
Does anyone know how one can get access to the ChatGPT4 model on platform.openai.com?
If there a way how to make GPT 4 smarter like it was before? I am ready to pay more, ready to pay per request etc but this gpt 4 stupid like old gpt 3... It do not remember conversations history, generate non workign code in 90% situation. Any advice?
api has the older model
GPT-4 API is available for all paying customers.
I do have a paid account and "pay as you go" to use the API but I do not have access to ChatGPT4 on the Playground
API you mean i need to create my own UI for this? 🙂
Or there is way to use it somewhere like in Playground
May I ask since when do you have it ?
What are all the options you see when you click on the Model options ?
I also have paid account but i dont have gpt 4 in playground.
The Plus subscription is separate from the "Pay as you go" plan
what do you mean "Pay as you go" ? There is some topic about it?
Do you still have the free trial grant ? Did you pay anything related to the API ?
Yes you can set a new payment method in the settings so you can use the API.
Oh!!! Really!!! I have paid plan on chat open ai but it looks i have free trial on playgorund. So i will pay per request if i add my card?
Yes but only after your free trial expires or completes...
How i can complete my trial right now? 🙂
If there a way to test old 4 version right now?
Only new account?
You can complete your trial quickly by asking too many questions in the playground and using the most expensive model. I guess it's 5$ ?
In my case, I already pay per request but for a short time. I imagine the GTP-4 model will be available later?
I don't know anything about the 'old' model.
From the picture it looks like you didn't use it at all this month. Did you use it the previous month ?
Right now chat open ai gpt 4 version is stupid. You said API has OLD gpt 4 model.
Nop
I didn't say the API has 'old' model. It has a gpt4 model
Sry, not you. This guy 🙂
In this case you are still not considered a paying customer since you haven't paid anything yet 😄
OK thanks for the info 🙂
playground or create an api key and get some chatgpt replica project from github or #1037561385070112779
Thanks!
Oh, now i see this.
Which version is smarter?
0314 is from march
It's probably the earlier one but You can test them with the same question and then compare. Which one is smarter depends on what you consider smart.
It is hard to say which version is smarter with one question.
Or may be i dont know the right one? 🙂 Do you have?
Yea that's a problem 😆 I guess it depends on your use case ?
I noticed that gpt 4 start being stupid after all day owrkign with it and fight for working code example...
Why do you think it's being stupid ? What did you try?
I have been using gpt 4 each day for dev for the last 4 months. I really sure it start being much more stupid. Not just me, my other dev friends also noticed it.
AI lost the point which we just talked with him few messages before. He start generating non working code. Etc etc...
I noticed that too but I am not sure if it's just about us being too used to it or is it truly becoming stupider. I haven't made accurate tests to compare the old vs new. If you still have a conversation from the old model then you can compare.
@heavy adder in the playground, you can set the 'system prompt'.
Oh i see, thanks!
Do you know why i got banned each time when try tro send some image related to the question? It is just one min but i cant understand for what i got it? 🙂
It's automod, it probably detected a non English language somewhere in the image. Sometimes it makes mistakes too.
You were not banned though, just shortly muted 😆
Gotcha, thanks 🙂
Do you know any way how the code in playground could be styles liek a code? Right now it use tripple quotes like like this but it not styled.
In the playground text area there is no styling.
ive been trying for hours now. how do you get whisper to transcribe audio in real time??
help
Are you using the API ?
yes
I used the local version of whisper, it takes a few seconds to transcribe after the pauses.
Is anyone here using the alpha version of ChatGPT? It's been giving me an error for the last two days.
how did you do it?
If there is silence for 2 seconds then the audio file gets sent to be transcribed by whisper
also i actually did not know there is a locally available version...
You can check the official GitHub repo of whisper
how does that work though? completely free no tokens?
Yes
wth...
Hello, there is a section called "metrics" in the legacy documentation. But there will be a new version in 2024. I couldn't find an answer to this in the new API, does anyone know?
How do I change what model I want to use in ChatGPT. I want to use the 16k version of ChatGPT 3 instead of the default.
Really new model of 4 version (gpt-4-0613) is more stupid then the old one...
Here is example from my work where AI totally frogot about array of formats for update function.
Why yes I do want to use that model. I mainly use the default version although the more tokens will definitely be better in some cases.
If there a way how i can see how much AI last answer cost in tokens or in $?
1 token is usually 1 word. Though it changes with contractions and hyphenated words. Specifically with English anyways.
How much a generation costs can vary on thousands of factors.
If you're asking if this is a feature on ChatGPT then it isn't unfortunately.
Does it means that settings in right size of length 256 impact on how smart the answer will be? Or it will just generate more long answers? 🙂
Oh you're using Playground?
yes, how you able to use oldest model in chat open ai? 🙂 Or you use some custom ui? 🙂
I don't ever use playground don't really know much of the gui. But there should be model selections.
Hello im new here, can somone guide me to what are the best prompts to use GPT 4 for research? Research would be more around a very specific health topic. Would like to know as much as possible about that topic to understand it as a regular person.
so how do you use old model?
I don't.
According to the gui there is a model button on the right of the screen just above temperature and maxium length.
The model you don't want it gpt 3.5 turbo. So try and find gpt 3.5 with that word.
@sinful mirage when you downloaded the offline version could you find C:\Users[Username].cache\whisper anywhere? i cant seem to find it
Yes the models are located in "C:\ ... \ .cache\whisper\ "
i had to make my own since i looked through everything in users and couldnt find it
hopefully this doesnt mess it up
I see, if you have any other issues then you can ask in the #1037561751362863144
Hello team, sorry if this is the right channel to discuss this issue, or already been asked by someone.
Seeing continuous errors throwing up as below, for Web browser
"Something went wrong. If this issue persists please contact us through our help center at help.openai.com." I did necessary troubleshooting steps. Can any one help me on this?
Me too with a premium account
I have a Premium Account. I use a 100 Mbits line to use ChatGpt. However, prompts are hardly answered during the day. I pay for the account and still get the "error message": "network error ". Today on 09/15/2023 only one prompt was answered without error message. What is the reason that I can no longer use ChatGpt in my Permium account as I could about 7 days ago? The constant error messages are brand new. I am not using any new devices or other software. I have checked the data line and I know that my computer has enough capacity to receive replies from ChatGpt. Please solve the problem, otherwise I will have to cancel the premium account.
your access to browsing is probably some unofficial extension?
Why does GPT-4 ask me to solve a puzzle every time I ask a question? It's very annoying!
Why is the ChatGPT orange and refusing to generate? Is ChatGPT offline? The History is refusing to load and I can't talk to it. It was fine moments ago.
Sorry for late response.
Orange ChatGPT might indicate that the servers are getting high load and the AI is not at maximum capacity. That means you're only talking to the API. (I think so)
prob should contact OpenAI help
Yes, thank you, but in the support chat they sent me here, and this is actually the first time I've logged in.
Question is this comprehensive
Question two is openai corrupted
Mocking jay 🐔 op vroom warp speed 🚄 openai op
You still tink you should ban people bc you are more elitest tan dem
Is there anyone from canada?
The message times in the json export of conversations are accurate, right? I just spent about an hour trying to get timestamps. Then I just exported all my data, found the timestamps, found a program that converts it to real time, and calculated it.
I thought OpenAI didn't save times of messages but I guess not.
Hello, I need help because I just bought the chatgpt+ yet I am not able to change the model nor am I able to insert a file, or turn on the web browsing function within the beta drop-down, since it isn't even there. I feel like I have been scammed out of my money.
Hi all. Here's a log in problem anyone can help? My friend log in Chat GPT 4 successfully everyday and this morning the log in page shows wrong email or password. The mailbox used to sign in also can not log in as well. Is there any other approch to get support and find back the chat GPT account? We have very valuable data in this account. Thanks a lot.
You are using an extension, they disable browsing time ago https://twitter.com/OpenAI/status/1676072388436594688
We've learned that ChatGPT's "Browse" beta can occasionally display content in ways we don't want, e.g. if a user specifically asks for a URL's full text, it may inadvertently fulfill this request. We are disabling Browse while we fix this—want to do right by content owners.
2929
431
Best place to buy 20k chatgpt?
Does anyone know if you ask chatGPT problem solving maths questions that are considered at a late high school to early university level, will it output incorrect answers?
For example, I have found a consistent pattern of 25 + 73 = 101 or 9 x 7 = 77 when it’s supposed to be 97 and 63 respectively.
Of course this is an oversimplified example just to illustrate what chatGPT is doing wrong but just wondering is this kind of scenarios possible?
since i used a second google account for sign-up here they said that the api access was disabled. was that for both accounts or could i still use the old one?
I signed up I dont know when , but I didn't get any credits
is that related to my posting?
the requirement with the openai account seems relatively new
no
My ChatGPT has been continuously used by unknown outsiders. Even after changing the password, it is still being invaded and used. What should I do?
why am i getting Error: 429, You exceeded your current quota, please check your plan and billing details.? here is my code:
import requests
import json
def run():
prompt = "Once upon a time, "
api_key = ''
url = "https://api.openai.com/v1/engines/davinci/completions"
headers = {
"Authorization": f"Bearer {api_key}",
"Content-Type": "application/json"
}
payload = {
"prompt": prompt,
"max_tokens": 50
}
response = requests.post(url, headers=headers, data=json.dumps(payload))
if response.status_code == 200:
response_text = response.json()['choices'][0]['text'].strip()
return response_text
else:
return f"Error: {response.status_code}, {response.json()['error']['message']}"
previous month i reached soft limit but this month nowhere close
Same problem here. Can't use it since 7 days because i get error message in 9/10 times. Also will cancel plus subscription, why pay for "premium usage" while you can't even use it at all?
hmmmm, it seems that openai is having issue ? I cant use it
the networks keeps throwing 403s, i even tried logout and in again
oh, this is stupid, i cant use a vpn ?
sir, read what other people are saying before you ask a question
you have to upgrade and you will be able to use it with no glitches, FIRST READ AND GOOGLE before asking a question
what do you mean, i paid for it every month
well, no matter the case, i turned off my vpn, it works now. so im good now i guess, just a bit annoying
you ran out of trial or your prepurchased credits
vpn is allowed from supported regions but with vpn you are using it with others and there are usage limits above user accounts
why is that, why do i have to deposit, i remember it was every month charge?
what does it say here for you https://platform.openai.com/account/billing/overview
monthly billing works like before unless your card was declined or etc.
sorry if this is the wrong place to ask. But I here a lot of people talking about using AI for note taking in downloaded lectures. Whats the best way I can do to this for free? please if you can give me a website link/name. Also, If there is nothing, does anybody have something that can turn my lecture downloaded video just into words?
you can use whisperai to convert the video to words, it's free but you will need to run it locally, which might take a whilehttps://github.com/openai/whisper @sharp pewter
thanks
can i install this in fedora?
as long as you can run python on a machine, it should run
i just tried and my system doesnt have enough ram
chatGPt is not working for me, anyone having the same issue?
For the last 24hrs I have been unable to use chatgpt. I am a paying customer with a paid current account. I have tried logging in and out and different browsers. I keep getting "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." with every query. Anyone help?
Try clearing cache and a different browser
Over intrusive antivirus, browser extensions and vpns can also cause that
They periodically seem to open some testing to some users. I have no idea how the people offered access is chosen. Sometimes it's opt-in, sometimes it's 'you were chosen'. Sometimes it's a small-scale pilot, sometimes it's many or all plus users.
Congrats, it appears you were quietly given access, on purpose I hope, to alpha testing at least for now.
There could be something special about your use case, I don't think any of us know.
It appears you can choose to use it, or not.
That 'lock image' on ChatGPT-4 is also unusual, at least my account doesn't show a lock on mine.
that's the old alpha menu from the time before plugins were released to beta, it looks like you have glitched ui
Thanks, will give that a go
Help.
How do i fix this? my text size is at default in my web browser same with my zoom level.
i cant read plugin descriptions or view the website properly at 100% zoom level
Hi - It looks like my account is being used by someone else, I can dozens of separate chats. I've change my password but it's still going on. Is there a way to kick them out of the account?
I dont have any specific answers for you, but you should definitely scan your PC for any malware or viruses, change your email password and add 2FA if its available for your mail service. I'd start there. Wish I could be of more help
And sorry for the late reply
Thanks - Tried all that. I don't see an option for 2FA on OpenAI, would love to use that. Do we know if there's any way to reset all cnx so that the other user(s) gets kicked?
I don't know of any, I'll ask the rest of the team
I believe 2FA is currently disabled for OpenAI, but I was recommending 2FA if you have Gmail or anything
Why is there no method of canceling a Chat-GPT subscription?
@jagged surge Also I would recommend removing any browser extensions you dont trust.
Hi
I have a question. If I add a new member to my plus subscription does that new member also have access to the plus membership? If I make them an owner?
I have a subscription but haven't been able to access ChatGPT for over 2 weeks now. I get the same error message over and over again
ive got the chatgpt 4 subscription but for some reason i cant use the same credit card to buy a payment plan to use the api... any ideas?
claude isn't an openAI product
Haha true might get in trouble in this server aye
You're free to discuss and ask about non-OpenAI products and services over at #ai-discussions!
i just got the same 2 days ago even tho i was using it everyday all day long before that for months
It just happened like that and impossible to make it work since
its learning... growing...

Guys hi, I have a question. Long time ago I accidentally deactivated my Chat GPT account and I wish to bring it back. Is there a way to get it deactivated?
activated again*
Once you delete an account, it can be reinstated. The email and phone number attached to the deleted account also can't be reused
i genuinely believe the site is getting worse, double prompt submissions, input prompt changing to what i think might have been another users somehow, responses that are completely unrelated to the prompt, resubmission of the original prompt when asking a follow up question, this is like the complete opposite of a Bethesda game, it was working fine at the start now its just broke
i guess my question is, why it seems like none of these issues are being fixed
are you using unofficial extensions
nope
is there a support chat somewhere?
original prompt was "whats the key difference's between C# and F#?" changed to "Show me a code snippet of a website's sticky header in CSS and JavaScript."
im just interested in knowing why some of these issues have been around for awhile without any mention / attempt of fixing them
I'm having some trouble using chatgpt, it keeps on giving me the "Something went wrong." error, any help? I've tried a bunch of things
Also it works fine on other devices that are on the same account for some reason idk
disable nordvpn web protection or try firefox
oh i see it was nord ❤️
Run a boot-time scan and then deep scan. Do you have an antivirus application installed? We're not in a stone era.
tried all that (i don't use a VPN also) and it didn't resolve ^^
i was told that we can use our own gpt api keys in openai's playground but i don't see the option to. what are our best options for using a larger context window in gpt 4 than the one provided on chat.openai.com?
if you're using OpenAI's playground, it automatically utilizes your API key because you'll be logged into your account. The API key is primarily for people making or using third party services that utilize OpenAI technology
Is Anyone else Experiencing Having to do puzzles when prompting ? for some reasoning it is contentiously making me do them
?
does anyone know if chatgpt-4 on the website (not API) has a context window of 4k or 8k tokens?
anyone else notice Browse with Bing is now live on the iOS app but not on desktop for some reason?
I am consistently having this issue although I have chatgpt plus.... Any thoughts, comments?
8k context window and 4k max_tokens value. This basically means that overall length is 8k and single request limit is 4k.
Hello I have a quick question, what's the best way to learn a new language with chat gpt, in a way that he can continues to teach me the next day in the same chat, without losing any progress, and he can track my progress. Idk if it's possible but I would love to know
does anyone know how to use chagpt api key on ios without being hacked?
in app you made?
yes
.
you need to have your own backend server which does the api calls and the app uses your backend
how do i do that
you need to pay for some cloud server or setup one some way and then run your made backend app there
i have no experience of making backend tho 😦 😦
you can ask chatgpt to help or find google articles for example about nodejs backend apps for custom gpt apps
Chat gpt through any browser shows unable to load history and prompts me to upgrade to plus even though I am already a subscriber. This only happens during the morning, in the evening I have no problem with access, any ideas?
The mobile version works fine.
try with firefox
Worked, Ty
The API outputs everything all at once, but ChatGPT outputs bit by bit with the blinking thingy. Is that just an animation, or does ChatGPT actually work differently?
you need to use streaming, it's not an animation, its returning the tokens while its generating em
having the entire response all at once is surely easier to handle in code
but streaming allows you to display the content sooner
j
I dont know how to fix this? can anyone help
it happens with every question and ive restarted multiple times
test with firefox
Every time you close the app or go off it the chat history goes away
Look in the very, very top left of your chat window:
Looks like your screenshot may not show that portion.
Don’t have that button
Try to hover over that area.
I did
Okay. I have no ideas, except perhaps #1070006915414900886 it. Good luck!
Well, the general 'if something seems weird, try rebooting the computer'. That might be worth a try.
Is ChatGPT down? I can't access the website
Sorry, the problem was solved. but just now there was a warning that the system was busy, then the website was down for a few minutes. I hope I wrote this message in the right place
Happened to me too. Trying again solved the issue.
Hi, I need support... chinese people are chatting in my account, and deleting my chatgbt chats
why's playground's chat so weird? are there settings in it to make it more convenient to use like it is on chat.openai.com? i'm assuming there are benefits to this layout that are lost on me
Have a system message geared towards your needs and make the temperature 0.7
Playground is not intended to be a place for chatting but for testing the API for its applications
That's why ChatGPT exists
gotcha. i heard it can use our api key so i was trying to have something similar to chatgpt on there. thanks for the info
How does one prevent GPT from using the altered text formatting?
I have "do not use markdown" specified in my custom instructions which works pretty well
Though that gets rid of all markdown not just code boxes
Can you explain custom instructions to me?
This is still the first conversation with GPT I’ve had.
Custom instructions allow you to specify special information you want ChatGPT to know and how you want it to respond
Should be somewhere in settings
Custom instructions are generally followed very closely
Thanks for your reply, I'm glad to know that it's a general system failure.
You should avoid using visual descriptions because it can't understand. It does not know what the typed message looks like on the screen. You can try sending the following message to the ChatGPT. It will probably give an inaccurate description.
"In your last message, you gave a character template example in a markdown box. This box separated the template example from the rest of your answer and improved readability. Can you give a visual description of the markdown box in your last message? Is your template example separated in a rectangular frame? Is it in a circular frame? Is it not in a frame? If it is in a frame, is the background color of the area inside the frame the same as the background color of the area outside the frame? Does the text in the template example you sent contain a different font than the text in the rest of the message? Did you italicize or bold the text in the template example?"
ps: If you want to ask for a visual description of a message, you should ask for the last message it sent. Generally it will not understand descriptions such as "the message three paragraphs above the last message" or "the fourth paragraph in the last message."
Does anyone here know plug ins well? I have a idea but only better then layman's terms knowledge.
I know what a API is but no clue to create and implement it
i was playing around with python and the dalle openai api. I wrote a python script to generate an image, download the image, use the downloaded image as the source of a new image variation. I ran it for 200 iterations. After 200 iterations, the image looked absolutely nothing like the original. I wanted basically to keep the original theme so that I could string together the images as an animation sequence. There is no way to control the temperature in the python method openai.Image.create_variation. How can i generate an image sequence whilst maintaining the original theme?
I tried also using masks, but i found the documentation on masks was not very detailed.
Here is the code I used to generate the sequence
#!/usr/bin/env python3
import os
import openai
# Define the size of the image and the animation length
animation_length = 200
#
PROMPT="""A man stealing copper from a 5 star hotel pool while drinking a beer.
He always faces towards the camera, wearing a red-tshirt, and has white skin,
he has ugly teeth, he looks malnourished. The style is realistic photo. Show the whole scene from a wide angle/"""
# Generate an initial image
response = openai.Image.create(
prompt=PROMPT,
n=1,
size="1024x1024"
)
image_url = response['data'][0]['url']
image_urls = [image_url]
# Now mask_files is a list of filenames like ['mask_0.png', 'mask_1.png', ...]
# Edit the initial image for the remaining iterations
for i in range(animation_length-1):
os.system(f'curl -o image_{i}.png "' + image_urls[i] +'"')
response = openai.Image.create_variation(
image=open(f"image_{i}.png", "rb"),
n=1,
size="1024x1024"
)
# Update the image URL for the next iteration
image_url = response['data'][0]['url']
image_urls.append(image_url)
here is the animation. i liked the first few frames but then it went awol
You might need to use another library . Idk if this is a thing but you might be able to find an image equivalent of fuzzy wuzzy to determine if an image is close enough to the theme
Ultimately I feel like you’d probably want to manually review the images. Idk if full video auto generating is in the tech yet
Yeah so it's a noted behaviour of Dall-E that extensions/variations slowly lower the quality of generations
not really a way to from from A-B while maintaining consistency, currently
Hello,
I' have been hacked by a chinese guy.
I changed my password but he can still connect to my account and request some chats.
I dunno what to do, i cant set up 2 authentification and i cant block his IP adress.
Thanks! My next step maybe was to generate 4 iterations, then regenerate the first and start with another four, and so on. That way I won’t drift too far from the original prompt. Thanks for the advice all.
G
does anyone know if you can use an image in the prompt for dall E? you can with midjourney but im not sure
you cant with dall-e
How do I fix this
no, you can't
sadly
Help me
Is that not possible, for ChatGPT to send some request over api on secret phrase or so? He will only answer in text, right? Can't do any action?
Is that a tip for me or someone else
the API authenticaiton works based on APIKeys and the reply is always a sequenc eof tokens in plain text
never do requests on the side of the client with your APIkey, if you are making a service, the API requests need to be abstracted by a service made by you on your server side
Your message is too vague. It is unlikely that anyone would know how to answer it with so few details
you might want to edit it and provide more context
he got picture in his message
#1070242883677982771 message that's how you request help for a bug
Cause thr guy who sent picture : chromium or not ? Pc ? Iphone Android? Api ? Chatgpt 3.5 or 4
Not wrong channel but lack of information
What I linked to you it is to report bug or issue you can also here but like if you have issues give some information so we cna try to help
there is no problem in asking for help on this channel
regardless of the channel, it is a good idea to ask good questions to increase your chances of getting a reply
"How do I fix this" with an atatched picture conteining an error id without any more information is barely information enough to deduce what that might be
from this picture alone I can guess the platform is not the browser.. and that is it
would be nice if the person added some information like.. is it ChatGPT? which platform, what they where doing when the error popped up? how often it happens and how long has it happening? Have you done something to attempt to fix it? if yes, what?..
¯_(ツ)_/¯
Is this the proper channel to be in for support if I am having issues with ChatGPT 4.0?
If it isn't, my apologies. If it is, here goes:
I just upgraded to Chat GPT 4.0 yesterday. Since then, Chat GPT gives me this message:
"Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
The help desk suggests a number of fixes:
• Using an incognito browser: Done with no success.
• Using a different browser: Done with no success.
• Clearing cookies: Done with no success.
• Restarting the computer: Done with no success.
• Disabling VPN: Done with no success.
• Try connecting to a different network (I work from home, so I'm not sure how that would happen).
Disabling browser extensions: Done with no success.
The odd thing is that Chat GPT 4.0 works from my phone, and it works from my laptop, both of which are on the same wi-fi network. Also interesting is that when I ask a question from my laptop, and then refresh the page on my desktop, the interaction shows up. I can't figure this out to save my life. Any help would be sincerely appreciated. Thanks in advance.
I'm new here, by the way. I'm a business owner, using ChatGPT to write marketing copy and copy for two websites that I am building for my business. Have a great day, all.
Try Firefox as it’s a browser not based on Chromium
I'll do that right away. Thanks very much.
That worked. You're my favorite person of the day. True story.
Thank you very much and have a great day!
By the way, I had previously tried Microsoft Edge and that did not work, just for future reference.
@fresh sapphire Edge is a Chromium based browser
That's really good to know. Thanks very much.
Hi i have problem with ChatGPT on my computer! I can't use it there i get "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." what can i do to get it to work on my computer! it works normal on my phone
That happened to me a lot of times, learned it the hard way.
Hi Arwen. Try downloading Firefox. It worked for me just a half hour or so ago. I had the same issue as you are having right now.
See above. I'm just learning to use discord, too.
aah thanks will try firefox i normlay use Chrome
As do, but it appears there is something about Chrome that it doesn't like, although, like you, ChatGPT seems to work just fine on my laptop using Chrome and my phone using the Chrome browser. Maybe it just hates my desktop for some reason.
yeah it works now thank! yeah don't know why chrome is like that but i now got it to work thank again for the help
It says that everything I put something in that chat
On my computer it is says the conversion is to long how do I get around that
when u play for plus are you not suposed to have better response times and less downtime
with gpt 4 ive noticed increase in downtime and decrease in speed
Now I'm getting "Our systems have detected unusual activity from your system. Please try again later."
This is quickly getting old.
do you have nordvpn or avast, etc. installed?
I turned it off this morning. I resubmitted the question and it worked right away. The randomness of this is a little disconcerting. I hope they'll figure it out eventually.
That happened to me too... but only once
Is there a way I could upgrade how long conversations can last for free
Hello, is anyone able to offer some guidance or suggestions as to the best next steps for this issue? I received the following message from OpenAI: *"Thank you for reaching out to us. We’ve confirmed this action was associated with suspicious account activity. Please check that you are accessing OpenAI’s services from a supported country. If you are using a proxy service, it's possible that the system detected this as a potential trigger for security concerns."* I have confirmed that I have not used the platform in an unsupported country. I have no reason to believe I have breached any User Policies in my chats, which I can no longer access. I also cannot access via Chrome, Firefox, Edge, or the ChatGPT App. I am not connected to a VPN, I've tried deleting my cache & cookies, logging out, etc. The "bot" in the messages has stopped responding for almost 2 weeks now & I am getting extremely agitated as I use this every day! Any help would be greatly appreciated!
Help
Can you please clarify your question?
If your question is about length of a particular conversation, no there is no way to endlessly talk and have full context of the conversation, the best way to do this would be to ask it to summarize the conversation into a few paragraphs and then start a new conversation with it
I asked chatgpt
Have you tried using the service from a mobile device or a laptop? That's what finally worked for me.
it could have happened that your account where compromised, for good measure, change your password
and avoid using browser extentions, there are some reports of malicius extensions stealing credentials
Hi guys. Does your chatGPT works with VPN turned on? I'm currently in SF, in fact, OpenAi in 18th St is 3 blocks away from me. I can't use it with VPN on. Why do you think is that?
vpns do indeed can cause trouble
it is best to not use them
@rustic marlin oh ok. That sucks. I got the service to block trackers and ads. 😣
@rustic marlin I can use both service in Canada. Weird. What do you do?
Does gpt-3.5-turbo-instruct support function calling?
that happens due to the fact that VPNs will cram too many users into the same IP.. if a significant amount of users using that same IP are using ChatGPT at the same time, then, it will block the users
@rustic marlin that's a sound assumption
that is an empiric obsservation
although the devs at OAI have told already that they work on minimising the problem of VPNs on legit users
Is there an issue with Moderation currently?
they can't ignore the fact that ChatGPT scrappers rely on VPNs
Okay having sent a request to Moderation manually, the answer seems to be yes, it's sending garbage results back to me
Try dming modmail
Anyone have any idea when the browse with bing feature will be back?
So yesterday I was having a chit-chat with chatGPT for my assignment and it suddenly started generating two responses simultaneously, with similar context but different tone:
It happens rarely and I didn't see the related mentions in the release notes or the announcement; is this a new feature?
I have a question. On OpenAI's site, it advertises GPT-3.5-Turbo as having a maximum context size of 16k tokens. Is that able to be used by peons or do you need special permissions or do you need to give Mr. Altman $1b in Series ZZ funding? lol
It's available through the API
Yeah but like what model? Because base 3.5 is the 4k model, how do I specify I want the 16,000 token context?
cant seem to login to chatgpt. i am on paid subscription. tested with colleague next to me, he was able to logged in. where should i ask for support?
There is a separate model with the 16k context
gpt-3.5-turbo-16k
is there a Huge difference in Text 2Code generation from GPT 3.5 to GPT4?
Better results?
I had higher expectations, but not really noticed anything different in using either.
Thank you
Hello everybody, I seem to have some problem to enter on the website. It displays me this sentence :"Free Research Preview. ChatGPT may produce inaccurate information about people, places, or facts. ChatGPT August 3 Version" and a white screen unfortunely i can't access to the service how can i solve this ?
Hi, my chatGPT plus got hacked and I changed the password yesterday the person who hacked it still got access and all my previous history got lost! what should I do?
Contact support@openai.com
thank you very much
I get this whenever i try opening the website (started today) plz help
even though the status says its up...i get this
try with a fresh install of Firefox
does not work with a fresh install of firefox. I already contacted support@openai.com
hi ^^ does someone know when and at what price the enterprise version of ChatGPT-4 will come out? that one without limits?
same thing happens
you might need to contact support at https://help.openai.com
OpenAI Help Center
what the hell i going on ith ChatGPT?
the black screen thing?
why?
yeah
im getting it too
what the hell
idk whats up and idk how to use their ticket thing
screenshot?
missing CSS in the console?
wydm
press F12, go to "Network", hit reload (STRG+SHIFT+R) and look if some .css or .js file can't be loaded (red)
then FN+F12
and maybe switch the default
Network tab is next to Source
NEXT to 😄 so these arrows will show it
OH
i have a lot that doesnt load
there you have your problem 😄
so how do you fix it from there
probably hitting some cdn that is down
reload
what about mine?
could be... but dunno about "CERT_AUTHORITY_INVALID"
???
you need to reload
i get the same one and its most of them
FYI; I dont have any problems
i have no issue for any other site. and this happens on multiple machines , windows, mac, chrome, edge, firefox
lucky
so either its my isp (comcast) or i'm just stuck
seems the same problem... hit "console"
yeah its the same problem
something like this happens 2-3 a week
RR_CERT_AUTHORITY_INVALID
just wait it out
but
what if you clear your cache?
it said i should uhh
if you have timeout, you can just edit your last message 😄
I saw that message pop up and getting blocked 😄
i use chatgpt to make hundreds a day on a "q&a" site... but i just look at this as a break. so lol. i'll just go watch some videos or something until it works ago.
which browser do you use?
i'm pretty sure everyone that's down is hitting a cluster of servers out there that the cdn is down so it can't get the various js and css files needed to load the page
with safari you have a pretty rough time trying to delete your Browser cache
because its load balanced, that's why certain people can
google.
and if you are on a mobile device it might work because they are on a different cluster
just my thoughts
but it sayys its a cert problem... not a 404 or redirecting error
do you mean "chrome"?
i use chrome
fair, but like ALL these people? different os's , different browsers?
Hey All!
I can confirm that browser and OS is not important for this issue. Also a lot of people on the forum are reporting the same issue
What's the best channel for asking technical questions about things like fine-tuning, frozen models, soft prompts, etc.?
Yeah.
clearing cache is always the first step...
however:
I just don't think it's a CDN-Problem...
Because of the error you got..
But I'm not sure at all
its' all good bro @waxen river - it will work itself out
did my school do something like what.
just gotta wait it out. like i said this happens every week or every couple of weeks , something similiar
how long do you suposse this will last? im giving it a few hours tops.
you gotta think about what is on the back end and how much compute an operation like this takes. i'm surprised its up as much as it is and stable as long as it is
that's just 20 years of industry talking lol 😛
@crimson sluice usually its like an hour to a few hours
yeah, i thought so.
Also see the discussion on the forum. The issue might not be on the responsibility of openAI: https://community.openai.com/t/unable-to-log-into-chatgpt-online/384453/17
i'm pretty sure this is a cdn issue like i was mentioning. nothing we can do but wait it out.
Agree
i think i'm going to go take a nap. good luck all
Oddly, it works on my phone.....
Same here
VPN = (probably) other country = other connections = other servers = other responses
yeahh mine too.
would be kinda funny if this is a huge man in the middle attack 😄
Hahahaha man in the middle attack for https
does it matter if the mitm is at the perfect spot in the connection?
for example if you're an high administrative openAI co-worker? (or if the mitm is pretending to be an employee :D)
Yeah I know... If you think about it, it would be an "man at the end"-Attack 😛
I fixed it by going directly to the cdn url and telling firefox to accept the risk and continue to whitelist it
but it wasn't working for me, in the UK with no VPN
oh haha, im not gonna do that. Doesnt seem safe to me.
How to integrate chatgpt into video editing?
it cant do video editing. best you can do is ask it to write a python script that applies some simple transformation on a video file
Hello Guys, I am facing issue. I am using chatCompletion open ai api. But facing some issue when I try to make a call in my node js code.
Prompt:-
Answer the following questions and obey the following commands as best you can.
You have access to the following tools:
search: a search engine. useful for when you need to answer questions about current events. input should be a search query.
calculator: Useful for getting the result of a math expression. The input to this tool should be a valid mathematical expression that could be executed by a simple calculator.
You will receive a message from the human, then you should start a loop and do the following thing to generate user answer
Option 1: You use a tool to answer the question.
For this, you should use the following format:
Thought: you should always think about what to do
Action: the action to take, should be one of [search]
Action Input: the input to the action
Observation: the result of the action
Option 2: You respond to the human.
For this, you should use the following format:
Thought: I now know the final answer
Final Answer: the final answer to the original input question
Begin!
Question:- What was the high temperature in SF yesterday in Fahrenheit?
When I use this system level prompt using chat completion open ai api in postman. It works fine and generate output like below
Thought: I need to find the high temperature in San Francisco yesterday in Fahrenheit.\nAction: search\nAction Input: "high temperature San Francisco yesterday in Fahrenheit"\nObservation: I found that the high temperature in San Francisco yesterday was 72 degrees Fahrenheit.
But when I use node js to make a call
const chatCompletionAPI = async (systemPromptTemplate) =>
await fetch("https://api.openai.com/v1/chat/completions", {
method: "POST",
headers: {
"Content-Type": "application/json",
Authorization: "Bearer " + process.env.OPENAI_API_KEY,
},
body: JSON.stringify({
model: "gpt-3.5-turbo",
messages: [
{
"role": "system",
"content": systemPromptTemplate
},
...converstation
],
temperature: 0.7,
stream: false
}),
})
.then((res) => res.json())
.then((res) => res.choices[0].text)
.then((res) => {
console.log("\x1b[91m" + systemPromptTemplate + "\x1b[0m");
console.log("\x1b[92m" + res + "\x1b[0m");
return res;
});
I get response from above call as undefined. Now both call are exactly same. One is directly called using postman and other is called node js, but getting error when made call using node js. It is pretty wierd.
Its stuck in this things for about 20questions now. What should I do?
start a new thread
I have a lot of information in that chat
you could also just edit a prompt at the beginning of the loop to "never enter it"
Edit:
dont promt "never enter it" 😛
I meant:
just edit the beginning of the loop, so you dont enter the loop in the first place
there is no choices.text, https://platform.openai.com/docs/api-reference/chat/object
Summarize conversations every like 8-10 messages then use the new summarization as a prompt for a new chat
Can I do that even if it says it is to long
Like the conversation is too long or one message length is too long?
It stops responding saying it is to long
What do I when the chat is already at the limit