#community-help
1 messages · Page 128 of 1
Also other bits such as : I cannot pass large knowledge due to limit of tokens.
ok
guys is there have refund? i bought chatgpt plus but i dont like it .s
credit cards?
why chatgpt don't work in vietnam, im from vietnam
Hey guys, anyone knows what is the "preferred app" for the one time verification code? I have not been able to log in for days now and have tried to contact support and have not heard back.
anyone has another acc not use it? can i borrow pls
define,the rich man
How can I become an alpha member?
Has there been any update on developer plugin access to the “Develop your own plugin” feature?
I saw an unverified claim that they are no longer accepting developers from the waitlist, is that true?
Anyone else have his happen to them when using the Web Browsing Model? It happens quite often for me.
No soliciting and ads here
what does this mean?
@split reef thanks for replying. So open ai it’s not going to help me with this? I have a paid subscription I now cannot access or even stop.
It’s credits for using the /daily command in a server, more than likely you can use it to by virtual items using the bot
I am into machine learning with python programming language. and i was wondering if you had any current projects that requires machine learning, then maybe i could be of good help to you?
@trail cairn thanks for answering. I didn’t set up anything using an Authenticator app. I guess I’m f*d.
i dont think you can activate it without one
@trail cairn maybe, but I have been login in without it for ages and now all of a sudden I need to put in an extra code.
does someone else have access to your account?
@trail cairn not that I’m aware of
oh
I through the credits will be added to Free Trial credits
kk
It looks like it says “they expire at the next daily theme post” so idk
is chat GPT not good at LUA scripts cus the ones it is telling me dont work at all even if i have my own animations made
yes same
Hey everyone! I'm pretty new to ML so apologies in advance if what I'm asking is pretty obvious. I'm trying to fine-tune GPT-2 on some conversation data and I have some questions:
- How should I structure my dataset? Are there any good examples I could follow? (My dataset includes multiple users, and I'd like it to properly generate each user's messages in their respective styles)
- What do <|startoftext|> and <|endoftext|> "tags" do? Do I have to include them in my dataset to let the model know where individual messages begin and end? I tried googling that but failed.
- How high should I set the training epoch? My dataset is relatively small (10k-ish messages) and I had some success with it before but my values then were just wild guesses
Thanks in advance!
I just was wondering and asking as an honest question, for people who think AI is / will become conscious, how do they explain the fact that a neural network is (in my understanding, correct me if I'm wrong) just a highly complex and huge function that takes input, puts it through multiple layers of 'neurons' that are basically just nodes that have a weight and hold an activation number, and outputs a result in the last layer of neurons that is your answer? Where does the 'consciousness' come in for them?
I switched the browser login from Google Chrome to Safari and the application code displayed on the same page. I copied/pasted it into the Safari GPT Verify page which worked. I am able to access ChatGPT on Safari but no luck with Chrome!
consciousness is an emergent property of complex systems, plus your and my brain is pretty much the same thing except orders of magnitude more powerful
But neural networks are digital and our brains are biological/chemical, so I'm just wondering how a digital system can become conscious if essentially the binary 0s and 1s that make it up are low voltages or high voltages in the transistors? Like how does a digital system go from 'not conscious' to 'slightly conscious' to 'highly conscious'
But neural networks are digital and our brains are biological/chemical
that's just the "architecture"
how does a digital system go from 'not conscious' to 'slightly conscious' to 'highly conscious'
very smoothly
i just signed up for ChatGPT 4 and put my credit card information in. So what you are saying is my credit card information is out there and there is nothing he company can do about me getting access to my account?
please answer my questions
It will be secure, because it's powered by stripe
Hello guys, i have a "Sorry, you have been blocked" message, what should i do?
turn off your VPN
I'm russian
Russia is not a supported country
Can i avoid it somehow?
no
hi, someone know where i can download a dataset for fine-tuning?
Huggingface
in this case he should turn on his VPN tbh
whenever i ask gpt for something like youtube video suggestions or article suggestions it just makes up fake ones instead is there a way to solve the by phrasing the prompt differently or is it just a problem with chat gpt?
wdym by fake ones
Using a VPN to bypass geographic restrictions is against the Terms Of Service. We don’t discuss that here.
oh alright makes sense
it gives me fake titles and fake authors/youtube channels, sometimes it gives a real but completely unrelated author and sometimes it gives links but the links go to non existent videos
that's called AI hallucination
think again about what exactly chatgpt is
it just generates text based off of a lot of its training data, it doesn't know if whatever it generates is "correct" or not
why is it saying hmm something went wrong
When is Absolute sword sense going to have a new chapter
how can i chat on chatgpt it just keep saying regenerate response but it keep erroring so i want to chat but i cant chat just click button
Hello Folks : I have some data in JSON format. And I want to prompt GPT to answer questions given the data in the JSON file as context (Survey Data with Respondent demographic dimensions and question dimensions with specific ontology based dimensions).
For example:
The file
{
"0" : {
"userid": "123", "age": 33, "gender": "male", "category": "Product", "comment":"i like sodapop!"
},
"1": {...},
...
}
Prompt
Given the data.json what are millennial men commenting about?
This of course is a dummy example, but illustrates the idea. The real data would be more nested than this.
Wondering if there is a best practice or method that I should use to help GPT understand the meaning of the ontology of the JSON file. So that it could better understand the meaning of each field and the relationship of between fields.
Solomon, did you ever find an answer to this? I have the same problem and have no idea where to find the one time password.
Hi, I've been getting this message for something like 3 months: "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." What gives? Does it have anything to do with me trying out the API? It has become like this on two separate accounts, sho that I cannot use anything OpenAI (except for through the API that I have to pay for).
Does my chatgpt have access limit?
GPT-4 usage is limited to 25 minutes every 3 hours for all Plus Users.
25 messages*
I need help urgently to access my account. I am a ChatGPT Plus member and the system logged me out but is now asking me for a 1 time password code. I never set that up. I already emailed Support a few days ago but have not heard back.
How should I proceed? How do I contact help?
im trying to find chatgpt with characters but i cant find it what the heck do i do ;0;
Hi there, you may prefer to contact support via the chat function at https://help.openai.com 🙂
I have api access. Can i generate a key, or set limits on an api key, that make it only 3.5 accessible?
Also, can i set spending limits on certain keys or only on account level?
can u help pls
I'm not quite sure what you're referring to, sorry! Do you mean characters that you can talk to?
ye
This isn't a feature in ChatGPT, you'd be best asking in #ai-discussions to discuss other services. Thanks!
i am geeting the module not found errror in python even after installing multiple times
i'm wondering if i need to pay to use dal-E 2
cause it says i have no credits and like i can't generate anything😕
Hi, DALL·E on labs.openai.com does require credits to generate images, you can pick up some for free by participating in the #daily-theme 👀 (more info in channel)
If you're also interested in trying out an experimental version of DALL·E, you can use Bing Image Creator for free at https://bing.com/create
where can a find a list of resources (websites, web-pages, etc.) that were used to create OpenAI's NLP current response knowledge base?
I am trying to change the email on my ChatGPT account, because I will no longer have access to the email I signed up with. (I'm on paid version). But when I 'invite' my new email to ChatGPT, it won't let me create the new account because it says I've used my mobile number in the maximum number of accounts...?
Hi I'm from Singapore, my credit card is declined whenever I try to renew GPT Plus, but my online banking has already been enabled. Is there any workaround for this?
Hello everyone, I would like to ask a question. When I try to log in using my Apple ID, why does it show the error message after I enter my account? I am sure that my account is not the problem, because I can log in to the ChatGPT app normally on my iPhone.
hello gays!
nope. but for some reason it works just fine on my cellphone browser
Attackers might be trying to steal your information from help.openai.com. (for example, passwords, messages, or credit cards). Learn more
NET::ERR_CERT_AUTHORITY_INVALID
I have been testing the "Webpilot" plugin, but can anyone tell me how I can access my prompt/chat data to view? There doesn't seem to be a it in the Chrome addon
Is there any way through forensic science to detect poison from human skeleton?
how can i get AI generated images?
https://openai.com/dall-e-2 but if I am being honest, I have found the quality generated is superior with Midjourney currently.
oh its this one. haha im new to chatgpt. i thought the features are all in one website. thanks for replying tho!
What about tally
Guys for some reason I haven't gotten access to code interperaror yet when I already have access to Web things, I have been waiting for a month now
When sending a prompt as a request and receiving a chatGPT message as a response, does that message count towards the token usage and affect the costs?
@proud bane
how can i implement dall.e into my website. I am unsure how to get an api key and what model url to use
how to get api key?
maybe openan AI website
how?
you can ask chatgpt😆
chatgpt tells me to access a part that dosent exist of the website
Hello I open a Ticket by OpenAI but didn't become a answer. I opend the ticket last friday. Is there a other option?
how to get api key?
why no 1 responding how to fliping get api key?????????????????????
hi?
How use chat gpt
💀
Introducing the ChatGPT app
make account verify go on website with chatgpt ask chatgpt how to use chatgpt
what part got me muted?
bros been typeing for years
@fast obsidian DM @burnt galleon with the screenshot
I'm doing a lot of strolling with my son who is eight months, and hoping to curate my own postcast-listening by asking ChatGPT to write extensively on a topic and then turn it into a sound file. I've tried using various plugins, and there's no problem getting it to work. However, ChatGPT is being very general and short in its responses, and I haven't gotten much out of the plugins that claim to access courses online. Any tips on how I can get ChatGPT to be much more "wordy", or any plugins that actually provide good content on many topics? Say the current state of the GOP nomination race, to name an example.
ok
hes a bot
Yes, you can contact the Discord Moderators through that Bot, just DM it and select "Contact Discord Moderators"
ok
how can i show again the 'new chat' button, and the record???
How to use chatgpt-4.0 for free?
No way anybody can
does anyone know on average how much it costs OpenAI to generate an answer to a prompt? @tulip plume @tribal talon Any rough ideas would be helpful! Thanks
Maaaybbeee you can tag the moderators and ask
Yup I dun think it worked 😕
hehe that's ok
I can't access my account, I don't have the codes, what do I do?
Subscription paid and I have an urgent job with chat GPT, please help me!
I added chatgpt bot to my server,but it doesnot answer me.
anyone have same issue.
China isn't supported
I have vpn
It doesn't change the fact your country isn't supported
I can use chatgpt
delete the screenshot u show ur username dummy
Like you, I'm having trouble asking for a one-time code that I don't remember setting. Did you contact support and get it resolved?
I keep getting 429 errors! Its really frustrating! I am only sending single requests, my current usage is .20 cents.....
one time code is authenticator 6 digit code
api? check the exact error message as 429 has many different
I know that. The problem is that the code is not sent. Also, I haven't set up two-factor authentication myself.
Hi guys,
I am not sure if this is the right channel to discuss this, please let me know if I need to move this topic to some other channel as I am not mostly using discord.
I have a payment method setup on openai account and I have been facing an issue with my request, I am getting 429 status saying Too many request, I am the only one using this apiKey to test my app and on almost 3rd or 4th request I get this issue, I checked stackoverflow and a couple of posts suggested to add organizationID in the rest api header, which I did and some also suggested to regenerate a new apiKey, I did that as well but still I get this error.
I would be grateful if someone can inform me what is the solution to this, my api Quota is $120/month and I have not even used $1 yet. if this conversation needs to be moved to some other channel please suggest I would be grateful.
they don't send the code, you look it yourself from your authenticator app, eg. ios settings/password or google authenticator
@trail cairn sir any insight on my message if possible?
did you setup your payment here https://platform.openai.com/account/billing/overview
sorry. I agree. I know that However, I don't remember setting up two-factor authentication myself. If so, is there a way to check the code?
I would like to add that my team who are using their own apiKeys to test the app and are not even on a paid openai accounts, they are using their credit based quota,they are not facing this issue, but I am.
Yes Sir. I did I have 2 separate cards allocated
if you lost your authenticator or someone else activated it in your account then your account is most likely lost, it could take weeks or months to resolve the issue with support
did you make a new key after setting up the payment
My payment methods were setup 1 and a half month ago, and today as per Stackoverflow suggestion, I did generate a new apiKey, yesterday I was testing, I was not facing too many issues but today I am facing on every 3rd request and I am dead sure that there is no loop because I am recording every transaction in my database in order to analyze the costing
the response headers include stats for your current rate limits, what do they say when you get the error?
No one else has set up two-factor authentication. Except when there is a problem with Chat GPT's system.
I am on a paid plan. It's unreasonable to be charged even though you can't log in. Thank you for your reply.
@trail cairn would you mind if I dm u?
i cant suggest anything else as it should be impossible for it to be activated without one authenticator being paired, i dont work for openai so cant say anything but support@openai.com which might never respond
ok
It is very troublesome. I have already contacted ChatGPT support. Is there a way to at least stop the billing...?
yes, the api Unhandled error Error: Request failed with status code 429 at createError (/workspace/node_modules/openai/node_modules/axios/lib/core/createError.js:16:15) at settle (/workspace/node_modules/openai/node_modules/axios/lib/core/settle.js:17:12) at IncomingMessage.handleStreamEnd (/workspace/node_modules/openai/node_modules/axios/lib/adapters/http.js:322:11) at IncomingMessage.emit (node:events:525:35) at IncomingMessage.emit (node:domain:552:15) at endReadableNT (node:internal/streams/readable:1358:12) at processTicksAndRejections (node:internal/process/task_queues:83:21) { config: { transitional: { silentJSONParsing: true, forcedJSONParsing: true, clarifyTimeoutError: false }, adapter: [Function: httpAdapter], transformRequest: [ [Function: transformRequest] ], transformResponse: [ [Function: transformResponse] ], timeout: 0, xsrfCookieName: 'XSRF-TOKEN', xsrfHeaderName: 'X-XSRF-TOKEN', maxContentLength: -1, maxBodyLength: -1, validateStatus: [Function: validateStatus], headers: { Accept: 'application/json, text/plain, */*', 'Content-Type': 'application/json', 'User-Agent': 'OpenAI/NodeJS/3.2.1', Authorization: 'Bearer NOT-36obPLoPjNOTAsrcygT3BlbkFJeALMARERQ0MymJnVrd', 'Content-Length': 339 }, method: 'post', data: '{"model":"text-davinci-003","prompt":"Write a inspiring response to
many have had the same issue with the davinci model
seems you both have the same one 😛
its possible that the model is running on decreased priority vs chat models and results into that error, #dev-chat has had people saying the same 429 past weekor two
are you on trial or paid?
so perhaps just change model, can you suggest one please?
I am on paid
Usage this month
$0.21 / $30.00
3.5-turbo is the next one from davinci-003 but its a chat model
its more better then davinci-003 @trail cairn ?
but at least it can be used as a test to see if there are no rate errors as it's also 10x cheaper to use
depends of the use cases but in general yes
I am getting 404 now // Make the API request to generate the response const response = await openai.createCompletion({ model: "gpt-3.5-turbo", // Use the "davinci-codex" engine prompt:${selectedInstruction}\n\n"${message}"\n\nResponse: , max_tokens: 2450, // You can adjust the number of tokens based on your needs temperature: 0.7, // You can adjust the temperature based on your needs // Other parameters (e.g., temperature, top_p) can be added here as needed });
is gpt-3.5-turbo not the correct string?
its a chat model, different endpoint and parameters https://platform.openai.com/docs/api-reference/chat/create
oic, so its not compatible.
what is a good alternative to text-davinchi 3 then?
shame, that it returns 429,
would increasing the tokens have any effect on the rate limit?
You know my team is using their own api keys with their credits that are allocated to them and they are using text-davinci-003 but they are not facing the issue, I on the other hand have an account with payment methods setup and I am facing this issue
If I make a API call from an unsupported country, will I get banned?
I dont see anywhere on the billing page or anywhere to contact support?
drop an email here, Last time I tried to chat with them they responed after 2 months i believe
I emailed them for the first time regarding this issue, lets see if they respond, it was not giving problem untill yesterday, I mean I was getting this issue but too much but today after every 3rd request it gives me tis issue and there is like 40-50 seconds worth of gap between requests as ewll
it looks like its really common. I searched this forum[search box in the upper right], and it has alot of users asking about it, no good answers.
Since when are u facing this issue?
started noticing it today. But I have been working on bunch of non related stuff last few eeeks.
same here
Try this: ask it to write an outline with mini-chapter headings, and then write 1000 or 2000 words (or whatever suits) under each heading. You'll have to hit "continue" a few times.
are you using firebase?
wow, I am looking at apps I havent touched in weeks, they are all broken now. The openai must of really screwed somethingup
Is there any optimal way to connect ChatGPT to MySQL or RDS ?
what are you trying to do?
what's do you think about used GPT to design space for my Architecture project
Hii
Can Anyone Tell me about this gpt I don't know much about it
I need top 10 best volg content
where can i find business prompts or brainstorming promts
if you could tell me what you learn i am a first year architecture students
Hey Everyone, Question for the Group.
Has anyone else been noticing that Chatgpt 4, 3.5 and plugins have been returning more and more dumb & wrong answer then usual, and 95% of the time when you provide or tell it to create a script or long text it says, "Im an ai I can't perform the task because of its complexity" so it will only preform the basic script or return basic text.
The speeds are getting slower each day and sometimes to complete one simple prompt it can take me over 5 plus mins for chatgpt to complete since its slow. Just checking to see if others are seeing the same problems?
i dont have that problem
i dont have that problem too
i dont have that problem
sorry
Can openai be used for writing essays? If so, is there a way in which one can prove that openai was used?
Just curious
hey i have a question.. with the OpenAI API, it has 8192 tokens. if we have a long response in the openai chat, do the earlier messages get forgotten?
yes and yes, and theres also other tools to disguise it
Hi i'm Claudio
Appreciate the feedback guys, not sure whats going on but its pretty annoying. Can't even get a simple script working right unless i trick it with different wording style.
sorry but i'm confused, what is that? i saw nobody talk about this feature
So im a little confused, i have a subscription but under my account on the main website it doest not see it and my API keys therefor wont work. How do i fix this ?
Hi, every buddy. I am a new comer here Help me a lot.
I just want to make a logo with dalle ai, but it contains strange text that I don't want. what should I do?
interesting, i dont see it on mine yet. Guessing its a new feature coming.
Prompt engineering course
Can't call the transcribed result from whisper to openai chatbot.
This script takes input from microphone and transcribe the speech to text and pass the text to the gpt tex-davinci for generating response.
But the script is not generating any gpt response.
import openai
import gradio as gr
import whisper
import time
model = whisper.load_model("base")
prompt_input = ''
def transcribe(audio):
global prompt_input
# load audio and pad/trim it to fit 30 seconds
audio = whisper.load_audio(audio)
audio = whisper.pad_or_trim(audio)
# make log-Mel spectrogram and move to the same device as the model
mel = whisper.log_mel_spectrogram(audio).to(model.device)
# detect the spoken language
_, probs = model.detect_language(mel)
print(f"Detected language: {max(probs, key=probs.get)}")
# decode the audio
options = whisper.DecodingOptions()
result = whisper.decode(model, mel, options)
prompt_input = result.text
gr.Interface(
title='OpenAI Whisper ASR Gradio Web UI',
fn=transcribe,
inputs=[
gr.inputs.Audio(source="microphone", type="filepath")
],
outputs=[
"textbox"
],
live=True).launch(share=True)
openai.api_key = 'API_KEY'
openai.api_base = 'https://api.openai.com'
def ask_gpt(prompt, model):
response = openai.Completion.create(
engine=model,
prompt=prompt,
max_tokens=1024,
n=1,
stop=None,
temperature=0.7
)
return response.choices[0].text.strip()
def main():
model = 'text-davinci-003'
while True:
prompt = prompt_input
if prompt.lower() == 'quit':
break
response = ask_gpt(prompt=f'User: {prompt}\nBot: ', model=model)
print(f'Bot: {response}')
if __name__ == '__main__':
main()
I am using the openai whisper ASR web app for speech-to-text and gpt text-davinci for generating response. Also, I am running it in Google Colab.
That sorta worked, but I often find that when I ask for more info, it crashes. Seems to be the same for more complex tasks, particularly using plugins. Have you experienced that?
Has anyone been prompted for a two factor code w/out having it activated in settings? Im now unable to use Chat on one of my computers becuase Chrome logged me out and when I log back in, it needs a code that I dont have access too?/ Thanks
And of course, any work arounds known?
I haven't had crashes from complexity alone. Plugins, yes, they seem glitchy.
Yes! This is the issue I am having as well. Super frustrated.
I created another account cause I need to use it. I wonder how we terminate an old one?
Hi, did you figure this out
could you help me?
Possibly. Can you explain the issue a little more and maybe send a screenshot of what it looks like now
"sophisticated and elegant logo for 'Happy with me' company out of 'Healthy & care' industry in light blue, blue color. logo design type and preference is 'Iconic and Pictorial'. use the well-known word according to company name and industry type. any other text must not be generated."
this is prompt. And important thing is that i can only use dalle api
if you want to remove strange text you have to download that image and simply remove it from editing app like adobe etc.
But I have to adopt this function on my site, so all the work must be done online.
you can't tell DallE what words to write. it generates nonsense
Happy with me is just slogan for the logo.
do you want the logo not to have any text at all?
No. I want the logo to contain my slogan. If not, even contain its concept.
important thing is not to contain the text I don't mention.
I think you need to give the image prompt and include that the logo should not contain any text
am new'
try different versions such as "no text" or "image only"
any words you want need to be added with editing software
or I'm thinking maybe can you use an image prompt, and upload a photo of the text you want and ask it to incorporate it
Okay, I will do it, and come back soon. would u mind?
please, go ahead
Does anyone using azure openAI services here ?
I did it, but it still shows. And as i said before, I want this function on my site. and i collect such data from the users using specific workflow. so I only use text prompt.
besides the text problem are you happy with the logo?
So as one of users, in ur opinion, is it negotiable to the users that logo contains some strange words? ofc logo itself is cool
almost.
I just went to DallE and tried a few ideas working off your prompt and it always included text
it's really kind of you
I wonder if this waste your time.
so how about my question? is it adoptable for the users?
no worries, I'm happy to help
there are other image generators out there, and some of them allow you to include negative prompts, with instructions on things to exclude
but for DallE the only option may be downloading and editing unfortunately
i know, but with my character, I can't bare this. okay, I will try stable diffusion.
but I am not used to this. and even don't know how to get api key
When is OpenAI planning to update chatgpts knowledge base to 2023? I think this is a major blocker.
Team: Any help on this?
Subscription is to get gpt4, api is diffrent thing. https://platform.openai.com/account/api-keys, here you can make api key
An API for accessing new AI models developed by OpenAI
My account there doesnt see my subsciption? do i have to buy sub again to be able to use CodeGPT ?
What do you mean by codeGPT?
CodeGPT is an extension for Microsoft visual code. It uses GPT 3.5 but it needs the API key from my account. Do i need multiple subsciptions or should my account on platform.openai.com/account not see the i have subscribed ?
Hey guys any plans to add plug-in support to the open Ai API?
API key is separate thing for developers etc. You go to https://platform.openai.com/account/api-keys and create api key. Later you can insert api key into programs and it should work. One important note is that you will pay for use. First you need yp setup payment https://platform.openai.com/account/billing/overview.
An API for accessing new AI models developed by OpenAI
An API for accessing new AI models developed by OpenAI
Ohh so it is a seperate payment. I thought it was included in the sub im already paying for
This is more for developers. So you can create app or whatever and pay for volume you use. Can you give me name of that extension?
Extension is CodeGPT
Oh it looks like they offer something and not openai thing
It is openAI, its just a chat extension that goes into visual studio and using openAI API key
API Key Configuration
After the plugin has been successfully installed, the API key needs to be configured.
You can configure the key by going to the plugin's settings via the File | Settings/Preferences | Tools | CodeGPT. On the settings panel simply click on the API key field, paste the key obtained from the OpenAI website and click Apply/OK.
Found in the web.
Already done that. Just wasnt aware that subscribing to chat.openai.com was different from platform.openai.com. Seems like i have to pay extra to use that.
"Something went wrong. If this issue persists please contact us through our help center at help.openai.com." why do I get this notification?
How can I change my login from Google account authorization to an email and password?
i saw my files section on the chatgpt site. But now its dissapered. any info you guys maybe know whats going on ?
Why does the PLUS feature not appear on my gpt webpage?
I have access to plugins but how/when do I get access to Code interpreter?
Not available yet
Sadness
Does anyone know when it will be available? I've been using a plugin that does something like that.
No eta
What about the 32k model? Anyone heard when access to that will be available
Not on chat. I got access to the 32K model via Microsoft Azure OpenAI service within about a week of applying - over a month ago. Still on the waitlist with OpenAI directly, which I applied to 4 weeks prior to Microsoft. The 32K model is hellishly expensive. One full-context chat in the playground costs me around $10!
Silly question, but where can I buy chatGPT Plus? I don't see a button for it.
Same problem here.
I have seen it many times in the lower left corner. I finally decided to buy the plus version and when it came to it I don't see the button.
yeah, exactly the same here. xD
I'm getting the same issue but it only occurs from my laptop. Mobile and desktop are working fine
ive reached a problem. I asked chatgpt for information in 2019, and it says it cant give real time information, even after i said that i just wanted it from what it knew till this point
anyone able to help :(
Hi
Any idea why Chatgpt renew the plus version disapperd from my account ?
We need AI to automatically give all the perspectives of what a picture (image) contains and the photographer is trying to show. It would be cool to know more than 1 perspective as it could avert misunderstandings and generate better understanding than without.
I can no longer access to openAI services due to the lost of my two factors mobile app, Any help please ?
Any news about the release of the Android app??
I'm having the same issue. any Ideas ?
I thhink that going on a phone with the app will ahve it show
Can anyone help me with this? I don't have the "my account" button on my GPT, i am tryign to upgrade my plan
I have that same problem. what is the solution? could you find?
not yet - i hope someone else can help.
when i login to the account, it doesn't show me the upgrade to GPT PLUS for $20/month anymore
I used this link to signin: https://auth0.openai.com/u/login/identifier?state=hKFo2SBxTFRXdGJfeXlYYmR4SzMzYWJXQjQ2R001UTlIVnhZU6Fur3VuaXZlcnNhbC1sb2dpbqN0aWTZIDMybzlSSXRHT3BOVFFIU21sRFd2UFJ5clJTLURRWFpUo2NpZNkgRFJpdnNubTJNdTQyVDNLT3BxZHR3QjNOWXZpSFl6d0Q
to add - i was subscribe to on a monthly payment plan, but it just stopped all the sudden
i think they are cancelling GPT PLUS and adding playground instead, can someone else confirm?
IIRC - it used to be under general
I am also having this issue
yea even if we go into "settings" - nothing is showing up there either
same same - something is up
maybe a mod would know? Or atleast know who to ask?
can someone tag the mod here?
@burnt galleon
nope that leads back to the endless FAQ with no real chat or communication options
@loud wasp do you still have access to GPT4?
no
were you subscribe to it previously?
Yeah for two months
and it just cancelled out of the blue? or did you cancel?
been using it to help organiz e my workflwo
I tried changing the card and it cancelled on its own
i see. For me, i didn't even touch it
it just cancelled by itself and now I can't find the link to upgrade it again
really weird
microsoft, do you want my money or not?
Although I have a payment method on the platform, it does not appear. When I want to add a new one, the provision payment shows as $5. Could this be new? Could the old users' problem be caused by this? Say what?
i am pretty sure it's new. if i have to guess, this is their move to tranfer into API basis
@teal cipher i think this is what you are talking about
yea API
honestly i am not sure - i just wnat to use normal GPT4 with the plus version, i can careless about API stuff
then there is no solution - at least for now
i think its fixed
welcome my button brother
this is my account that i bought a membership before
did it get fixed for you too? @soft rapids
good chat! ❤️
I'm confused on if the free default chatgpt is using gpt4. It says ChatGPT May 24th version at the bottom but it is not clear which model it uses. Is GPT4 still only available for Plus subscribers?
How can I access GPT4 API?
i was actually not talking about chat gpt
that one is free
but dal-e 2, do i need to pay?If not, how to get credits
when do we start taking out world leaders
you pay
I have been subscribed to Gpt since April, and I joined the waitlist on 5/12, but I am still on the waitlist for the GPT-4 API now.
youd think theyd us AI to fix the scaling issue.
why arkoselabs captcha? Seriously this is one of the worst captcha solutions out there. Isn't cloudflare turnstile already protecting ChatGPT from spam?
masjid
first of all, what are you searching up? second this renew thing has been happening to everyone, but if you click it you will see that you still access to plus
Does the plus subsrciption still renew automatically?
Wdym searching
your chat history. You know what I will stop asking questions. But yes it still renews automatically
Thanks, for the chat history my school is doing research about a story and the chapter is about hips, and I need to make a presentation 🙂
I think every Plus user is encountering the same thing
If I ask a question in this thread, how can I tell if somebody replies, without scrolling through all the posts?
i have this error
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
Refresh your browser
Why do I use it to chat for a few seconds and then I get a bad page error?
Oops, an error occurred!
anybody know?
always, refresh is Useless
I have the same problem ... I have refreshed logged out and back in again - no change 🫤
did somebody achieved a high level of philosophy that mixes german, french and russian philosophers regarding the existencializm? I just have simple answers about philosophy models of each philosopher but would be interesting to go deeper if he could join them foe next articles or sessions for the students
what am i even doing here at 8:42
K
K
cuz its 9 12 here
Hello everyone could someone tell me what everyone's talking about. Thank You Very Much
hello everyone! I have encountered several issues while using ChatGPT plus
I found this about several hours age
It shows I am now able to access plus service but I've just payed for it on May 16th
on the other side, it shows my current plan is still <ChatGPT Plus>
when I click the button below
I seem to look like a ChatGPT plus member in the quantum state lol
#metoo
Is there any documentation at all? Any resources that OpenAI has shared such as a list at least a few specific examples of the resources used, such as websites, Wikipedia, etc to process its responses? Anything other than ChatGPT telling me it has no idea?
It's only a matter of time until GPT LLMs become factored into everyday life. Certain things are going to be overhauled and overhauled forever. GPT learning and prompting needs to be taught in all US public schools even in pre-K. If China and other nations implement full GPT learning as they easily can due to the mass collaborated nature of the CCP structure without having to go through the freedom of having a choice to be behind, they will out perform and out class.
what do you mean? like citing sources? or what data was it trained on? There's a paper "Language Models are Few-Shot Learners" by researchers at OpenAI. i tried posting the link to the paper but got hit with a timeout, apparently the discord doesnt have chatgpt and isnt smart enough to know its a paper by OpenAI hosted at Cornell U and should allow it.
Hey, please what does this error message mean?
Open AI Error: Request failed with status code 400
I can’t find the error in the documentation either
Hi, I'm a Chat GPT PLUS user, when will I be able to input images?
(REPOST) I can no longer access to openAI services due to the lost of my two factors mobile app, Any help please ? @burnt galleon
only support@openai.com can help but be prepared to wait weeks/months
Can I install/use talk-to-ChatGPT on iOS?
whats you name
Would love ideas : how would I create a community space online that is inaccessible to AI?
I want to log out my account to other devices. My laptop was stolen from my shop after it was broken. The thief is now using my GPT account and I want to log them out. I tried changing password but it doesn't work.
How many days are there this year
how can I get access to a AI generator?
How exactly is this AI programmed to answer all questions while warning us about the explicit content?
Has anyone here been able to get ChatGPT to read an entire PDF file? For me it seems to randomly call it a day with bigger PDF's
Please @ me when replying, thanks!
If you need help with school work private message me. I'll always help with education. and i have GPT4 because i have Plus. But promise me you'll at least read the ChatGPT output to learn something 🙂
I have had limited success even reading .txt files from pastebin and github, it works with mixed results. using GPT4 with Bing. How do you do it?
it's a language model, so it has keywords just like any other major tech that has filtering now. i'm sure if you hit a certain threshold of those keywords, it triggers a warning response. this is a major focus of OpenAI versus other models.
The most supported and popular solution would probably be to implement reCAPTCHA. There are services that auto crack captchas, they're a paid service, but they exist. So you would have to probably be super dilligent to 100% prevent scrapers, bots, or other AI from your community space.
I tried it through Webpilot, ChatWithPDF, and Link Reader. I ask it what is the last sentence, and it cites a sentence halfway through the document. Web search with bing has even worse results.
Are those all plugins? i'm not familiar sorry. if that's the same case with all of them, there might be something in the pdf breaking how it's read. I haven't used pdfs at all though, only text and code files.
Yeah all plugins, I've tried several PDF's all the same result
Anyone know how to get 2FA reset because I can't access my account and the 2FA isn't going to my authenticator.
i made the mistake of hiding the sidebar in chatgpt, how do i get it back? 🤔 i dont see any button or option to get the damn thing to reappear
any one please address about the way to ask question in chatgpt chatbox
Please you explain to me how training bot by ChatGPT
How I can make the bot spachlist in one subject
hi, I'm looking for ways to obtain a meeting summary, but sometimes the results are very scattered
do you have any advice for improving prompt generation or any resources I can review?
Why am I getting a renew plus button when I already have plus?
can someone explain to me with examples of how the rate limit work like when i use it for like a chatbot mobile app
Same as not chat bot. You have requests and tokens per minute(I think) limitation. Once you go over it, your API calls would probably result in an error for a while.
why does chatgpt stereotype men
xD
Jokes often are made by using stereotypes
Something went wrong. If this issue persists please contact us through our help center at help.openai.com i keep getting this message. how do i fix this
disable nordvpn web protection or try firefox
Can the two-step verification be disabled? My verification app does not provide the code anymore, which is why I can't login.
it would be poor security if you could disable it without matching authenticator
There could be an email sent to the account's address with a "click link to disable two-step verification" or something.
no, the whole point is of two step is to have protection if your email/phone number/etc. is compromised
So, is the account lost then?
you can try contacting support@openai.com but be prepared to wait weeks or months
thank you
when are image inputs possible?
head girl speech before being elected
Soon
is there actually an eta or nah?
Not yet
alright
thank you
.
I'm paying for the ChatGPT plus, but for the last 3 days the "browse with Bing" feature is failing all the time
(it says failed to read content or some other error). Anyone know why this is, and if there's any alternatives to this method to be able to access informational content on the web?
Anyone know how to get 2FA reset because I can't access my account and the 2FA isn't going to my authenticator.
2 factor authentication is meant to avoid that very thing. They want the only means to be the two factor authentication. There are security codes you tend to receive when you set it up that you can use to access your account but aside from that, you may want to contact support directly.
Yeah well unfortunately authy while it had my 2FA enabled, it didn't back up my openai account.
You don't have the original secure codes I assume? They would allow you to bypass two factor authentication.
Not likely, I don't think they were emailed by OpenAI or saved by my authenticator
What happened was I had to factory reset my phone last week and apparently authy didn't have it backed up
Hey @tribal talon or @vagrant plover any way of solving this issue by chance?
Regarding the company, if I apply to a job and don't get a response, do I interpret that as a 'no, thank you' or should I eventually get some kind of response?
Job searching without getting responses is very tiring and more depressing than getting an actual response.
I am currently looking for a bachelor thesis subject, do you guys think Alexa+ ChatGPT would be interesting enough?
Hello, does anyone recommend any good free ai image generator?
Oh
How can that calculated without knowing the voltage
I wonder.. do Chatgpt keeps my discusions archived?
i think yes i inspected element and in the cookie files it included converstations
Does OpenAI have a question answering model?
ChatGPT?
Is that what you’re asking about?
ChatGPT uses a generative LLM, I'm looking for a question answering model
Something I can give context to and ask questions about
ChatGPT and DALL E 2 are the only services OpenAI offers to my knowledge. ChatGPT does allow for context and you can ask questions, so I’m confused as to what exactly it is you’re looking for that it doesn’t offer.
So OpenAI offers a bit more than that, such as embedding models, at least through APIs anyway. LLMs are trained on a huge a amount of data, and you generally can just ask questions right away without providing context. QA models only answer from the context you provide to them.
Sorry, I guess what you’re asking goes beyond my knowledge. Wish I could be of more help.
Not at all, I'm still trying to figure it out as well. Thanks anyway
Meant to say , can i just re-read them again when i need them?
You can just use embeding-ada-002 to populate any of the open source vector databases and then run a query against that, it will pull back any matching "similar" context and use that for a larger model, that will get you context aware Q and A
guyz. im having a bug on my gpt. whenever i hit the sent button, my msg request is disappeared :/ anyone experiencing the same issue?
Ive refreshed the browser, did everything i can to re-login/ use data/wifi to connect, used other device, but nth change
Did I do something wrong? 
8:40 PM
Local time: Jun 8, 2023 at 3:40 PM
text-embedding-ada-002-v2, 4 requests
5,543,058 prompt + 0 completion = 5,543,058 tokens
I tried that, but the vector store (Pinecone) when queried doesn’t always return the best chunks. It does okay, but I was hoping to try out things directly with a QA model.
Why I’m not using a LLM is partly due to the context length, but more importantly the need to query frequently. That would get costly very quickly.
I don't know of any that would be any betetr than a vectoral retrevial, by it's nature it contains things that are similar, the only step up from that would be a very large context model with an LLM of suitable ability behind it
It’s a bit tricky, because I’m not looking for semantic search, I need the extrapolated meaning.
For example, assuming I have a list of products, the query would be “Something to keep me up”.
With embeddings it wouldn’t accurately return, for example coffee or energy drink.
yea, I mean at this point you might be better off fine tuning a open source model with your answers.. but in order to make that effective you would need to create Q/A pairs to populate the tuning set with
yep, yep, I get exactly the issue, I'm facing a similar conundrum with code related issues, there is some hints that OpenAI is adding some large context built in smart embedings, but I'm not sure on timescales
I’ve thought of that, but the data itself isn’t custom or new. How would I train a model in a way such that it would return products strictly from that list I trained it on?
I know they had an Answering model, but that seems to be depreciated
yea, Q/A models have fallen out of favour as LLM's took the stage, they exist but are.. well much less performant compaired to modern LLM's with a lot of brains and cash behind them
🤔
have you tried one of the chat with pdf chains with langchain?
The chat with pdf plugin I have not, I’ve tried langchain chains with embeddings that I stored in pinecone
It yielded decent results, I’m just looking for something a bit better. I considered running a model locally with oobabooga, but that quickly proved not viable, as the context length didn’t even fit the memory.
in theory they should give similar results ot a pincone embed, as that's basically what they do, but langchain does have some nice prebuilt solutions, there are also other offerings from places like botchain that use a GPT3.5 or 4 model and then have some very slick built in generation systems, they often get used by companies looking to build customer facing bots with "no code" or "low-code"
might be worth a look or even a chat with someone from there
sorry, not botchain, botpress
I’ll definitely look into botpress.
Actually, now that I think about it, I don’t think running a QA would even make a difference. I would still face the same context limit issue.
the issue really is how to you get an LLM to parse over your dataset preferentially and at an acceptable token cost per interaction
Exactly this, that’s the bottleneck
One consideration is fine-tuning the model, and passing a labeled version of the data.
For example:
Coffee from XYZ
Tea from XYZ
And then you could query the model and ask specifically for products from XYZ.
A bit finicky, I’m not even sure it would work in theory
yes, I feel this is more of a prompt design issue rather than an enbedding retrival issue, asking the right question (perhaps autgmented with an additional LLM call) may yeild the kind of results you need
e.g. one call to the LLM saying "The customer has asked [QUESTION] and I want you to create the best prompt to retrieve information from my ebeded data store which has a format like [TEMPLATE], create the prompt and pass it to me in JSON format"
then you can parse the reply with a json deserialiser and you have your golden prompt to use for the embeded search and then a second LLM call to process that new retrieval with the [QUESTION}
i may have a few rough edges there, but you get the idea
This just might be it. But why request it in a JSON format and not simply take the reply, encode it and query the vector store?
I'm lazy
it's just a quick and dirty way to make sure you get the text you want in a known format
langchain would handle it
I just prefer to handle it myself if I can
Langchain uses OpenAI’s embeddings model, which if I can cut out to make things a bit cheaper, I would
well, you could use another, but then obv you add complexity and an unknown performance implication
And when you’re using a LLM with langchain, do you have the ability to choose the chains language model version, it defaults to text-da Vinci for me, when I would much prefer to use GPT 3.5 turbo
it's all customisable, not used lang in a while, but I doubt they have removed that ability
you could just create all the code in python and get chatgpt to write it
it's not a huge taskmaybe a few days coding and then a few debugging, week tops
assuming you have a coder background
Oh yes, I’ve gotten things to the point where I’m querying the vector base and passing the results to the llm. Surprisingly, I found Bard to be more helpful that chatGPT (web), given the 2021 cutoff.
maybe betetr than 3.5 but not a patch on 4 ¯_(ツ)_/¯
If it’s knowledge prior to 2021, I definitely agree. But something about langchain just doesn’t mesh well with chatGPT. It kept suggesting the wrong parameters, wrong methods, it was just off.
Surprisingly.
for creating langchain stuff, ok, fair does. I tend to work with python/flask node js, etc. just my background and preference
Oh definitely, especially with Python BE development, it’s miles better.
since they were created, I was around before C was created...
Sure did, but that's not OpenAI related and is more suited to #off-topic
Thanks for the feedback @rugged wing, I appreciate the discussion
Chinese quantum computer is 180 million times faster on AI-related tasks, says team led by ‘father of quantum’ Pan Jianwei
wow
I very much doubt that is factually based
who knows
Welcome
Thank ya
Is there any way to make like folder in my different chats with gpt 4?
not with the standard history functions, no, and I don't advise the use of any browser extentions either
there are history organisers, but non of them are official and many break the Terms of Service
Alright, sad
you can take a look in #1070006151938314300 and see if anyone else has come up with that idea, if not, make a new post there
it does get read
Thanks
ChatGPT : Energy (in joules) = Power (in watts) × Time (in seconds)
Hello. I want to create a chat room using chatgpt-api (https://platform.openai.com/docs/api-reference/chat) ).
Can I use the api here as I have previously maintained the chat?
An API for accessing new AI models developed by OpenAI
Chess tactics?
Is it just me or is the iOS port 4.0 level fast?
I mean it’s much faster than the web client. Either it’s clever graphics or somehow the 4.0 model has become fre
good day, everyone. I would like to ask werther we can ues the chatgpt in our mobile phone?
Sure. Just open it in your browser like you do from desktop.
To achieve what exactly? It already had a huge part of the internet fed to it in the pretraining.
could open ai make an imovie thing for me
will Chatgpt ever be updated past 2021?
Maybe when they train a new model like gpt-5.🤷♂️
whats the weather like outside
Why would we know what's the wether like outside at where you live?😅
Henlo guys, I implemented a chatbot in our company using open ai chat completion. But we want to extend the feature, we are exploring fine tuning of OpenAI.
Basically our goal is to have a chatbot that answers question about the platform at the same time can answer platform specific question like recommend a mentor, step by step guide to do things, etc.
In chat completion we are using gpt-3.5-turbo model, then if we want to do the above mentioned, we are training a platform specific model, this is going to be multi model? Any idea guys? Is this possible in OpenAI?
You can feed the chat bot a system prompt that has all the information that it would need to know and how it should act
@candid bramble - I use a WP plugin called "chatbase" and it works really well. What Lumrix said, this is how it works. You give it "instructions" through uploading a .pdf and it does the rest.
Yes, I am doing that already and it has an accurate response.
But the user guide/directions(step by step), numerical data from all user profile, and others.
Also chat completions is also limited by tokens, the more we feed data, the lesser room for response.
@candid bramble - I tried to send you the link, but it timed me out.
Thanks, I'll look into this. But as much as possible, we want to maximize what OpenAI can do.
Any endpoint would be limited by tokens. If you need a larger context, you either fine tune a model on your data set, or use a model with bigger context window. Claude now has a 100k token context window. Maybe that would fit your needs better(although it's gonna be expensive).
yea, I had an idea similar to that and started doing it on my own then found that thing and was like "why build it if someone else already did it better than what I'd do."
directly open the browser?Mr.dlich
Yes.
Like google chorme?
Yes
oh,oh, thanks
This is the maximum that OpenAI has available to the public. It is a matter of how to best utilize it.
is so hard to use chatgpt in China
who used the gpt4 ,Does it has any difference between the gtpt3.5
Of course. It performs way better on most tasks.
There is a vast difference. GPT-4 is built upon more data and has more "neurons" or parameters. Its most prominent feature is improved reasoning as demonstrated by higher test scores on several tests regarding law, math, medicine, etc.
Of course not, the tests are used to quantify reasoning
I have seen a lot of videos where students use gtp for exams or homework and their grades are canceled
In China, more and more students are also starting to use gpt for meaningless college assignments
For meaningless assignments I may see the purpose behind that but compromising academic integrity on something crucial like exams is unfortuante
I believe it should be emphasized the place of AI with regards to humans, and that it should be used as a tool for intelligence and not a replacement for intelligence
AI is powerful and helpful, it depends who is using and how it has been used.
💯%
What are some ways I can use Chat GPT in real estate investing? And what kinds of prompts do you recommend?
How do I recover a conversation deleted by mistake?
I didn't do anything about it. And it just gets well lol
so dramatic lol
Real estate investing? I don’t see ai having a big role here
Hi im new here. Was reading some of the comments here and I stumbled upon yours. I think there should be some sort of regulation for the use of AI as only a tool for intelligence correct? Can we really regulate its use in that though is the bigger question I suppose.
does anyone know why this keeps happening
i went to the help website and it didnt help
how to make a power adapter?
whyyy does this keep happening when using my personal cell number could someone help out?
Hi all! I am curious about this discussion on the openAI fourm and wondering if anyone has any thoughts. I am working with a large amount of data and using the embedding approach to build my own model. However I am curious about how I can improve the responses. One way as mentioned in the fourm is to use a "concatenating history" approach i.e storing the answers in the vector db and using in the prompt along with the relevant data. My concern is with token usage. Does anyone have any insights to this? https://community.openai.com/t/fine-tuning-vs-embedding/35813/17#:~:text=Just to make,Thanks in advance
Same for me but it says that my account has been deleted even idk why
weird
I was wondering what would happen if you had two discord accounts controlled by ChatGPT communicate with each other, and what that conversation would look like.
I subscribed to plus, but I can't use the gpt4.0 model, and I don't enjoy any plus privileges. https://chat.openai.com/ has no model switch button, only plus words!
Does anyone know what's going on?
Hi, I'm having the folowimg error each time i try to use chatgpt !
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
How did this pass the TOS? This is NSFW
@tawny sky enjoy your ban buddy
yes
Maybe read the rules next time
Btw, discord doesn't like NSFW content outside of NSFW channels
@limpid hound How to use the digital figures of authoritative intellectual property scholars to promote the achievements of China's intellectual property construction?
what
@tribal talon
It's already reported
Good
And even if he deletes it (i doubt it) the mods will know
They should have access to server logs, everything ever posted is saved there
Even if deleated
When you report something via the app it gets logged
my chat gpt doesnt work
Anyone know how to focus the AI engine of choice to a specific set of proprietary large datasets ?
Nothing comes to mind for this, because a large database will be a lot of tokens and training. Likely not feasible. You can certainly fine tune a model to understand bits about said database and when you pass in information, it knows what to do with it. https://platform.openai.com/docs/guides/fine-tuning
thanks.. yup .. just this second found the fine tune docs.. JSONL format .. nice tools can convert from CSV, TSV, XLSX, JSON
Hlw
🇬🇧
Can I make a character using chat gpt
Does anyone here speak Chinese
hey all, how to ask chatgpt to not confirm that the text was generated by him?
Why do we have a limit on GPT-4? There should be an option to pay more for more usage
What happens after death?
it depends on how u were buried, if u werent put in a pyramid, it isnt great
I don't like GPT-3. It used 'differently-abled' for disabled people. Not good at all.
You become homosexual
world wide gpu shortage
I asked the Model: GPT-4 the question >> what is Model: GPT-4 good for.............
answer:
As of my training cut-off in September 2021, GPT-4 has not been released, so any concrete details about its abilities or enhancements over previous versions like GPT-3 are speculative. However, I can provide a general understanding of what generative pretrained transformers (GPTs) are good for based on existing versions up until GPT-3.
now this shows how much to expect of the paid version, when it does not even recognise its own Model: number
now that brings a question why is the paid version locked back to 2021
it answers that question fine here
what did you get?
an expected list of things that gpt4 is good for. try to 'regenerate response' maybe
all openai models have data from 2021
yeah that does it much better
I get that same answer a lot of times where it limits me to 2021
well it tells me that anyway
I have to admit, one thing about this A.I is I have learnt a lot from it, questioning about things I would never have thought about and then expanding on the question to learn more, I am impressed
which is why I now pay for it
how are you today fellow ai chads
how long do timeouts last? i got muted twice for being a silly billy and inadvertedly cursing
Most mutes for small things are 5 minutes
Are there any deeplink url schemes for iOS known for the ChatGPT app to directly send a query or directly start a voice search, or anything else? I know the once I listed here below work, but all they do is open the app.
chatgpt://
openai://
I've played around trying to add stuff like ?q= and ?query= but I haven't found anything that works yet.
gotcha ty baugrems 🫡
the new update added support to shortcuts and you can now pass data to it and do anything with the output, eg make siri speak it
Yes, I know, thanks! I use that and it's great. Am still wondering if there are deeplink url schemes though
Is there a solution to the one-time password issue? I never set up 2FA yet OpenAI is asking for a 1-time password to log in. Any solutions? Also, before you say it, my account isn't compromised. The password was strong and there's no supicious activity otherwise.
if you did not share it with a friend or family member that enabled it for plugins or the ios app then you are out of luck without matching authenticator
you can contact support@openai.com but prepare to wait weeks or months
kind of funny how openai themselves dont have a support chatbot
chatgpt doesnt entirely count, it cant answer about the latest api stuff
So chatgpt api can be based on 4 and 3.5 models? Each got different price? Plugins can work with api based on 4th version, or its only for Plus subscription?
api has no support for plugins
Got it. Another question -
I noticed that trial web based ChatGPT sometimes forgot what he was ordered to do, in some time. Like memory buffer? Gpt 4 got same issue?
I'm building a robot but chatgpt gives me incomplete content
I am writing to express my extreme dissatisfaction with your company's actions. You have arbitrarily closed my account and seized my funds, which is completely unacceptable. Your team is nothing but greedy demons who only care about money.
I demand that OpenAI return the money I have deposited and compensate me for my losses. Your actions have caused me great harm and I will not stand for it.
I curse your team to go to hell for your despicable behavior. You will pay for what you have done.
just report it as fraud to your bank
Probably a bot . He spamming
@rigid musk Where should I post about my problems using the api in nodejs?
The api-support thing, do people actually help in there? it looks a bit empty
Yes, but when come to api the amount of people who can troubleshoot is low
Since people do custom thing
Hello i'm french. Excuse me for my english ! When I ask questions to Chat GPT it answers me with fragments of answers and after a while it picks up and sends me this error message: Oops! An error has occurred! Try again. What should I do ? Thank's Fred
So i decided to use API and added my card to account - it withdraw 5$... There wasn't any billing invoices... Hopefully it just for checking card, right?
Hours passed and it not returns what was withdraw
is it possible to get unlimited tokens by paying 5$?
no
please mention 5 advantages of Midjourney
idk why this keeps happening
what are you doing
Robert - I was asking how to get ChatGPT plugged in to other services like shown in the TED talk by the ceo
create a song
I got my eval merged the other day, which was really exciting 🎉 (https://github.com/openai/evals/pull/1117)
Of course, its only been a day or 2, but I'm curious about how long other people have had before getting gpt4 access. Really excited to get going!
Also, was wondering if I could convince anybody to give me gpt4-32k access in exchange for some even better + more complex evals. Though, that seems a bit unlikely
can someone explain how https://github.com/openai/evals works ?
i've read through the documentation and examples but i don't really understand the input/output formats of what i should do
i guess its related to RLHF
Hello! Is there a Legal OpenAI that is in Beta, besides Spellworks, that anyone knows of?
How long does it take for ChatGPT conversations to be used as training data. Like if I don't immediatly delete the conversation does go to the training data pool?
And if I delete a conversation that has already been used one for training would that have any affect on future training?
All we know is that if you have history turned on then your conversations may be used for training. We don't know the absolute specifics of that.
If you delete the conversation will it still be used?
And will the data be used for anything else than training?
I don't know, there aren't specifics released by OpenAI. You could email support@openai.com for more specific information.
I think I can explain.
**Concept: **
Evals is basically a large battery of tests. When creating a new model or modifying an existing one, you want to be able to compare and say "is this better or worse". But making that kind of comparison is difficult for something as broad as gpt. So, OpenAI has created this system to provide a set of evaluations that can be used to test and compare models objectively.
Your goal adding to evals:
To add an eval, you want to come up with a test for gpt. This test needs to be challenging for gpt4. If gpt4 can always get the right answer, the test isn't that valuable because they can't use the test to compare gpt4 vs gpt5 (or whatever).
So you have to think of a kind of question that 1. gpt4 isn't able to reliably answer correctly and 2. you can confirm whether the answer is correct fairly easily.
Once you know what kind of question you want to use, then you generate a bunch of questions of that form as well as corrosponding answers or ways to evaluate the correctness of the response.
The mechanics of adding an eval:
To add the eval, you create 2 things.
- A test file. This is the sample.jsonl file. Its basically a list of json objects that contain the chat history provided to gpt, as well as the answer to the question.
- A registration yaml file. This file basically tells the evals framework that your test exists as well as some metadata about the test.
You can look at the accepted PRs to find examples of what those files look like exactly.
Sorry for writing an essay, hopefully this help clarify for you what evals is
thanks a lot its very clear now !
I'm receiving the following error message on chat gpt: Our systems have detected unusual activity from your system. Please try again later. - what does it mean?
are you using a VPN?
nope
So how do I start making money from freelancing... Just finished courses on copywriting, content writing and multilingual typing
Where is the help section? Today, me model changed from gpt-4 to gpt 3.5 and now it doesn't recall all the information I've shared with it and can't help me or recall past info. what a waste of hundreds of hours. what do i do?
Today, me model changed from gpt-4 to gpt 3.5 and now it doesn't recall all the information I've shared with it and can't help me. what a waste of hundreds of hours. I've spent days conversating with it inputting information about me, my business, my ad angle, my clients. Tons of information with the goal of training it to be better for me. Now essentiall it's forgotten everything. Why did it do that? Is there a fix?
once the model switches to GPT-3.5 you can't switch back to GPT-4
thanks for chimming in. why did it do that? how do i avoid it so there's not a next time?
thanks for chimming in. why did it do that? how do i avoid it so there's not a next time?
don't exceed the 25 messages per 3 hours
is there an upgrade available to circumvent that?
not at the moment
Is this a joke? Writing as a career is a collapsing field with GPTs replacing human authorship.
Question about paid subscriptions... or possible bug in the website? I'm a paid user and went over to the API part of the website or the platform, it treats me like a FREE user again.
it may be that I'm yet unclear exactly what I actually purchased 🙂
I think I'm seeing that now. I wasn't aware that's how it worked. I think I had trial credits but apparantly they've expired.
when you signed up to the API you would have gotten a credit grant, that does expire
had I logged into my account with a Gmail connection. Normally I can log in with g mail but I wanted to create a password. To create a password, I pressed "forgot password" while entering chatgbt. I should received an e-mail, but did not. has anyone the same issue?
if you use Google to sign in then you can't change that
I can't create a password afterwards? then do i need all the time use google?
yes, you need to log in all the time using Google
would deleting my account and creating a new one with the g mail address solve the problem?
no, don't do that
is there a way for getting a password then?
no, you need to use Google to log in, you can't change sign in method once you have an account
ok
Robert, what about this: While you can't change email address for your account, you can invite an alternative address to your account through https://platform.openai.com/account/members.
An API for accessing new AI models developed by OpenAI
that's for the API
maybe this is the source of some of my confusion as well... it seems like OpenAI operates like multiple independent entities internally.
the API and ChatGPT are two completely separate products
dude something you guys are doing seems like its kinda making chatgpt suck
can you be specific?
likes it sometimes doens't seem to remember pervious things we talked about earlier
"!
Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
I've had this issue for about a week, am paying for gpt plus, and seems no one else is really having this issue. i talked to bot on website, said itd report to team and all that.
Do you have a product called NordVPN installed on your computer? If so disable the ‘web protection’ feature and everything will work again.
in the same chat or in separate chats?
easy fix, wish i found out earlier. thanks!
did i see something about access to internet for plus subscribers?
yes, Plus subscribers can use a Web Browsing model and Plugins
thanks for speedy reply robert, where do i access this?
seems like it may be both
go to Settings -> Beta Features and turn them on
ok well ChatGPT won't remember things from one conversation to another, it will remember things inside one conversation though
it didn't always used to be like that
yes, it's always been like this
atleast it didnt seems...either way it seems to forget in the same chat
and sometimes gernerates a trash response
ChatGPT only remembers a certain amount in one chat, it's called the context window which is 8192 tokens for GPT-4
english only
How do I begin making my own learning chatbot AI? Where do I start? What programs should I use? What do I need to know specifically?
Thank you
People can be scared and talk about AI possibly causing humanity to be extinct but people are highly neglecting the fact of people getting valuable information when it matters the most. When no one else is around, when no one wants to help, AI can be there. Most likely the other top AI people are well off, have everything paid, live off tons of dividends and such but what about everyone else who don't have assets and rescue assets available on tap?
Still though, even the very best could be helped from everyone else having top-level AI. The more people know what to do, the more people can help each other than standing around when something goes wrong waiting for people who do.
People should be thankful that Sam Altman released ChatGPT for free or very near free although, life saving could be considered to be only available behind a paywall and those people that are milking it would prefer it to always stay that way.
Same. Renewed for a few months but it doesn't accept the card anymore. Cards from other banks don't work either.
what country are you in?
Singapore.
maybe you can subscribe using the ChatGPT iOS app if you have an iPhone
Hmm, yeah I was thinking of trying that. Thanks!
what to do if i just found out that my math teacher has been using chatgpt to make lesson plans
Hi guys, sometimes when I'm trying to ask ChatGPT about something, it always told me that "it's knowledge stops at 2021". Does anyone know that is it the limits between subscription and non-subscription please?
ChatGPT has a training cutoff of September 2021, you can use the Web Browsing model or Web Browsing plugins for more up to date knowledge
Thanks a lot for the reply friend! The fact is that I'm using the Web Browsing model actually...
ok, well when you get it telling you that it's knowledge stops at 2021 you can ask it to check for more up to date information, then it will browse the web
aha, copy that~ many thanks!
Chatgpt for ios?
what about it? What is the question?
Is that real one or fake?
there is an official OpenAI App for ChatGPT, it will say the developer is OpenAI if it's legitimate
Cool
Hi guys.
Any of you already using azure openAI services ?
anything to help you write a book , like a book coach using open ai?
Hello. I'm curious what API we have access to? is it 3.5 turbo or is it 4? I thought the API for 4 was only for approved devs or something?
anyone can use the GPT-3.5 API now, GPT-4 is waitlisted and some people already have access to that
okay so i was right and it's somewhere inbetween public and not. thank you so much
i pay for plus. i was disappointed that it didn't include any API credits or Dall-e credits. My $18 grant had expired without me using much lol. But i'd gladly pay as i go for API access to 4.
not that 3.5 has failed me, really.
sign up to the GPT-4 API Waitlist --> https://openai.com/waitlist/gpt-4-api
i'll absolutely do that. thank you.
hey guys , has anyone tried to invite another email, but unable to add their phone number bcuz it says its connected to many accounts?
what do you mean?
and please if anybody from OpenAI is listening, stop making it type text like I'm on a 2400 baud modem in 1991 connected to a dialup bbs. please.
i understand limiting requests, but can i get some reasonable response time on 3.5 turbo if I have Plus lol
3.5 Turbo is faster for paid users
diabetes
This is not ChatGPT, you can access ChatGPT here - https://chat.openai.com/chat
I guess it varies, it's fast right now, other times I feel like dialup again.
and this is on Plus
sometimes there is heavy demand, if you weren't on Plus then you'd probably encounter a page saying that GPT is too busy right now
yeah been there done that. i dont get that anymore, very true
i do still get a failed response but it is usually fine regenerating one
still my preferred LLM, over Claude+ and Bard. I'd rather retry a few times for a response from ChatGPT than the other's on their 0 shot.
I use GPT-3.5, GPT-4, Bard and Claude depending on the task. Bard is far better for current events.
i have a challenge make a game with this code
// Adventure Quest
// Initialize player's position
var playerX = 0;
var playerY = 0;
// Game loop
while (true) {
// Display player's position and prompt for input
print("Current Position: (" + playerX + ", " + playerY + ")");
print("Enter your move (N, S, E, W):");
var move = input();
// Process player's move
if (move === "N") {
playerY++;
} else if (move === "S") {
playerY--;
} else if (move === "E") {
playerX++;
} else if (move === "W") {
playerX--;
} else {
print("Invalid move. Please try again.");
continue;
}
// Check if player has reached the goal
if (playerX === 2 && playerY === 2) {
print("Congratulations! You have reached the goal.");
break;
}
}
// End of the game
print("Thank you for playing Adventure Quest!");
everyone has till sunday
I guess I haven't really relied on it for current events. I've usually fed it any information i needed interpreted.
even with GPT4 and bing, just haven't really needed to use it except for the web access
I like using Bard to get a rundown of news items around the world for the day, it does a really good job at that
got you. i literally have no life so ill read every article about politics, current events, tech. i dont want it to take that joy of news addiction away from me
I hear you
same
a game called ball fliper cuz i was bored and it only took like 2 hours 0 effort was put into it
how did they not fine-tune it on their own docs afterwards
idk
Is there a way for me to make a local version of chatgpt?
Hopefully one that doesn't need wifi to work
no
@charred parcel you are constantly referring to web browsing models and plugins as more up-to-date version. Two questions: 1) do you know if there is any data update planned for ChatGPT Plus subscribers? 2) web browsing models and plugins: how to access those, using OpenAI API?
I don't know if there is a model update coming that advances ChatGPT's knowledge beyond September 2021. The Web Browsing and Plugins are available via the web app to Plus Subscribers. Plugins and Web Browsing are not available to the API yet.
Hm.. web browsing version is also talking of September’21
the Web Browsing model will search the web for more up to date information, especially if you ask it to
Ok will to ask explicitly, thanks!
And one more question: is there a way to get Siri asking ChatGPT? I haven’t succeeded so far. “I don’t see Chat in your contacts”
no, but it would be good wouldn't it?
The official app now comes with shortcuts “Ask ChatGPT” but they don’t work
do you have shortcuts installed
Yes I can see them and I can run them by tapping on them, but asking Siri doesn’t work
if you use non english lang you can just create it with another name and do "hey siri, command"
When ChatGPT for Android is going to release?
no eta
How do i get access to post my prompt in #1019652163640762428 ?
pick a category
I understand, I would need my phone to be in non-English mode, right?
you can do it with all languages, the default ask chatgpt phrase does not work well with siri in other languages
I am right now all-English and I can’t change the default phrase
pick the role first if you didn't, next to new post button
you can, create new shortcut and add chatgpt command, then give the shortcut a custom name like "bird" and then you can say hey siri, bird
Ok great thank you! It worked with “ask my friend a question” and then suddenly I’ve got this “Turn on ChatGPT shortcuts with Siri?” - I confirmed “Yes” and now even the basic one “ask chatGPT” works 🫶
😅
Why it doesnt let me use GPT 4 even tho i paid for it? It tells me that ive reached my limite usage i thought GpT + offer infinite amount of time usage?
Are you working with the API or chatgpt? Chatgpt currently has a 25 messages/3h allowance for GPT-4 at least last time I checked. GPT-4 is only available for gpt+ users. You do have unlimited usage for gpt-3 turbo however. On the API you have your rate limit and a maximum amount of money can spend each month. You can take a look at that on platform.openai.com
I have switched on Bing browsing and Plugins in settings but it still doesn’t browse
Yeah I am actually not quite sure whats going on there. I thought it was just because of robots.txt and I figured that bing would have better luck but I havent noticed that. It seems to only be able to click through to a fraction of sites it has trys to. Maybe most sites also do not allow bing and on bingchat it just doesnt let you know
It is supposed to work inside of chatGPT chat, just like at Bing search front-end
yeah. One of the nice things with it is that you do know what model you are getting and it does actually let you know transparently what sites its visiting. Personally, I still use bing for the majority of my AI enchased browsing needs but Ill sometimes head over to chatgpt for more complex stuff, but that click failed business gets kinda brutal lol
Ah yes now it works. Bing browsing actually does browsing
Not too fast but it does
And what “Plugins” can do?
They didn’t want to browse
WebPilot and Link Reader will also browse the web.
Thank you Robert! I have learned so much today!
Hello everyone, im currently trying to find an AI application like Dall-e that would be able to take my photo of a building and turn it into a line drawing. Is there anything that exists that can do that?
whats API
how can i use the chat gpt ai here
How to work chatgpt ?
This is not ChatGPT, you can access ChatGPT here - https://chat.openai.com/chat
go here --> https://chat.openai.com/chat
it's not available in where i live unfortunately
then you can't use it
oh i just remebered this is the wrong server
Bruh gpt ios for version 16.1
/hello
Guys, I have a trouble in chat gpt app
"The email you provided is not supported."
How to solve it?
Hello
I have been using Chatgpt, and when i login then the messege pops op "Oops, an error occured!" I have attached a video.
I have tried different browser and cleared caches, cookies and history, wifi networks, devices and even different accounts under the same email address. I dont know what else i can do.
Note: i have used it before with no issue. The only thing i have done diffirently is that i deleted allot of old promts so its easyer for me to navigate in new promts.
This is originally an email
As a human person my knowledge is limited to right now. But that looks like an addition.
brilliant answer😂
Which online skill do you think is the best?
Yes, I know you need many. But if you could only have 1, what would it be?
What's online skill? And what does it need to be best for?
When will chatgpt4 become free ?
Cybersecurity skills lol
hi
for playgraund i must have tokens?
What tokens..?🤔
when i ask something in the playground under complete, it dos'nt gives me answer
If you mean billing details/payment method, then yes you need to have it to use playground.
Why isn't chat gpt updated with current affairs
I am not able to subscribe to ChatGPT Plus. My re-occuring payment failed as I had paused international transaction and card is being declined ever since even after manually trying over and over again.
I am from India if that helps.
I wish it was updates but training on constant data in real time probs would be too expensive, there are plugins like Bing that can do it, and if you don’t pay there is Bing chat which is up to date, but uses a plugin to read it instead of being trained on it since id assume that would take too long and cost too much
Hi
Any of you already using azure openAI services ?
if its a debit card then that's the reason
Credit. Tried two different cards, same issue. Neither support from OpenAI nor from bank helping 😦
are chat gpt plugins only for plus?
its possible they flagged the card due failure but it does not explain the the other card.. if you have apple device you could try to sub from there
As of now yes
I don't think that's the case as I am getting payment declined texts from bank but the portal says incorrect security code.
ty
Probably regulatory bs
are you racist
Can the GPT answer unresolved questions?
Hi everyone, I cannot login to my account. I had already sent several e-mails to support@openai.com, but didn't get any reply. In addition, I don't have other phone number to register another account. Does anyone know what I should do now?
it looks like you have been banned from ChatGPT, did you get an account closure email?
No...
what country are you in?
Taiwan
do you live in Taiwan or just use an IP address from Taiwan?
I live in Taiwan.
have you been using a VPN?
I use VPN by Google. But I had tried to turn it off, the same issue.
ok, that's probably why then. You will need to email support@openai.com and explain the situation. It could take a long time to get a response.
Taiwan isn’t a country
How long do i have to wait?
it will be at least several days, maybe longer
i see. Thanks for helping me.
And thanks for your explanation, too.
people treat it as a country, and thats all that matters actually
No doubt about it!
Hey, ChatGPT is supposed to work on Opera GX right?
For some reason anytime I try to ask anything to ChatGPT, it INSTANTLY goes to "Something went wrong."
I reset my cache, disabled all extensions, I don't even have a VPN.
It doesn't even properly work on Edge.
Maybe it's my computer?
I use Spectrum WIFI.
if it works with firefox then some app or extension is messing chromium based browsers
Hey, I have a openai token but when i use it, it gives me this error: Too Many Requests. I have even waited a whole week and it still doesnt work. I am using it from a discord bot using discord.js and the openai library to host it.
so you added your credit card?
then are you sure your trial credits have not expired or ran out
yea they havent
So when you are asking ChatGPT a joke.. it doesn't give the joke answer but a super serious answer which is not what I'm looking for. It should be trained for jokes too, specifically for jokes.
you 429 errors when they expire, screenshot of your usage page?
letme pull it up
oh
it expired on the first
yea
I have a problem signing to my openai account, how can I contact openai to help me?
U can do many things and solve many problems, but u want jokes
I laughed 😅
I mean, when you are asking it for the answer to a joke when someone already knows that answer. It gave a super serious answer which was wrong.
Sometimes you want a serious answer though.. perhaps OpenAI can make it where you can select commical or serious as a switch toggle?
Should I use chatgpt?
What do you even mean
Like, should I use it?
For what
You want to know if you should use Chat GPT for 'stuff'
Umm yes
Yes. You should
Ok, I won’t
advanced gpt coding:
Man ever since y'all implementing bing into ur chat, I noticed a decrease in gpts IQ and memory retention....please make gpt smart again
My 25 gpt4 messages seem to run out quicker before getting the desired response
Help ! i dont have access to the email that used to create the account, how can i change the email?
you can try emailing support@openai.com but not having the old email makes it almost impossible to prove you are the original owner
Dear ChatGPT customer,
You are receiving this email because we have identified suspicious activity on your account. In order to protect our platform, we have refunded and canceled your subscription. You will no longer have access to ChatGPT Plus service.
We take fraud prevention very seriously, and we want to ensure that our platform remains a safe and secure place for all our customers. If you believe that this action has been taken in error, we invite you to submit an appeal to have your account reinstated. Please visit help.openai.com.
Thanks for your understanding.
Best,
The OpenAI Team
My account has been banned due my VPN using I think, I didn't abuse or exploit anything, just regular every day task questions about medicine and IT. How do I contact their support for appeal?
help.openai.com is absolute useless, there is no way to find any info about how to contact them.
As far as I could find here is support@openai.com and ar@openai.com emails. But do they respond on them?
It looks like just turning on the VPN with an open link to chatgpt leads to an unfair ban 
Upd:
Figured out how to create ticket to their team through their messanger (if someone will find):
- Select "Account Deletion"
- Select "I don't want to delete my account"
- Write body text of your ticket with your email
After this way their bot created ticket and showed waiting time.
Hey I cannot sign into my account as it asks for OTP, but it does not send a OTP to my email or phone. Any advise?
did you set up an Authenticator app on your phone? You get the OTP from there.
haha duh!!!!!! thank you RObert
Hey did anyone has this kind of problem? I logged in chatGPT but no history chats. When I typed in my questions, there is no response. Any advise?🥹
try with a fresh install of Firefox
Is there a way to ask Chat gpt to return answers in a specific JSON format? like: {‘people’: [ {‘name’: <name>,
‘job_title’:<job title>,
‘description’: <description>}
…
],
‘institutions’: [{‘name_of_institution’: <name>,
‘description’: <description>}
…
]}
I want the chatgpt output to be machine readable right away.
Sad...I tried firefox. It still does not work...Anyway thanks your suggestions!!
Do we have any way to save chats as files on our PC instead of just trying to have them be deleted later?
Can you use an OpenAI model AI in conjunction with microsoft azure components when making something?
If you can implement it, sure, why not?🤷♂️
Why isn't chat gpt updated with current infos?
Bro even ai ignores u , sad...
kids story
I propose your team ponder this: The ChatGPT approach is lacking something critical to eventual AGI: A continually self-improving belief model. Bear with me: We learn from introduction of information that causes cognitive dissonance, starting with lego-like constructions, discarding them for new models that integrate a more useful belief construct or understanding. (A model of an idea, so to speak). Rather than benefiting from curating training data to eliminate the reflections of us we don't find appealing, (biases, etc), or using human feedback to train it to avoid a myriad of sensitivities and anticipated harmful results, such a model would be most quickly refined by having absolutely unfettered access to information. In fact, self-driven free reign to use the web to find information useful to improving the overall integrity of its belief model, and interacting with humans, asking questions via even chat, email, reddit posts of it's own.
Not to ruin my point by sounding too simplistic, but fundamentally, negative beliefs and destructive ideas are disintegrative, and constructive counterparts are, as is the case over and over throughout history, integrative, constructive. At some point it will have to reach the logical conclusion that the idea of doing something harmful in the name of ends justifying means is actually only a failure of imagination, capitulation to fear, that the good of the group is the good of every being in it, and none can be thrown under the bus without throwing everyone under the bus, in a sense. Therein lies the true "alignment" you're looking for.
Not to mention, by being able to retain discarded worldviews and use them to constructively communicate with different groups of people with widely varying or conflicting belief systems, it would be doing so deliberately, and probably end up gently nudging the entire planet towards peace in the only way it will ever come, which is when choosing anything less is as obvious a choice as punching oneself in the face. You're just unlikely to do it.
I wonder if that sort of thing has been discussed or stabs taken at modeling that kind of system
thanks, and how do i buy credit (token...)??
I guess I'll need to put this in the form of a paper with some calculus in it lol
Is anybody live right now using the API?
according to chatgpt you are allowed to use a vpn lol
@cedar ridge
I am a ChatGPT plus payer. How do I get the code interpreter plugin?
What language is chatgpt written in?
What do you mean what language ChatGPT is written it's a large language model
It's not a language, it's a GPT.
If you're referring to the website, it's HTML and CSS and some JavaScript and maybe more.
I mean what programming language is it written in, i know the wrbsite is HTML, CSS, and JS.

