Terrific tutorial! Great job of explaining the economics - super important for high volume low value transactions (e.g. data classification). Calculator is also a great tool.
@@gptforwork Your extension is awesome! The only issue I have is the "Exceeded maximum execution time", when I want to generate website content (more than 500 words) in Google Sheets… Not in Google Doc. I assume it's not possible for your team to fix this issue. Right? Thanks
Its indeed mouth watering seeing GPT trading in all CEXs at the same time .. i mean this is giving, it's absolutely amazing 🥰 . So happy I was able to get hold of it as it got listed on MEXC few hours ago. This awesome gem is built to create a trillion-dollar data and artificial intelligence (AI) revolution. Ain't seeing non of this token doing it . Good luck to them 🙂
How would I take the 100,000 words my CHatGPT generated for a thesis and python code And put it back into ChatGPT to refine the thesis? I coded a UI to change max tokens... Is there any model that can take an input that high?
Hello! The concept of max_tokens is still applicable, but newer models have bigger max_token limits. Check out our model comparison page: gptforwork.com/tools/model-comparator
Hi, I need to create some short description, let's say "maxTokens=30" and I tried something like "stopSequence["."] because it cut all the time my description in the middle. "stopSequence" is not working. Can you help me out?
Hello, we don't expose the stopSequence parameter, where are you using it? In any case, the max tokens parameter is not a length guideline, it will cut the answer even if it is unfinished. If you want a short description, for example 30 words, you should put this in the prompt itself. I can see how this would be confusing though.
@@gptforwork Hi by adding "max 10 words" in the prompt, they still where some descriptions longer than 10 words. So I am searching for another solution.
@@maxcoutaz4291 there is no perfect solution. You probably have to insist a bit. For example "strictly under 10 words". But if the task is really difficult it won't obey. gpt-4 is much better at this, but also more expensive.
Please add a or show me the) simple way for me to confirm which model is being accessed. I signed up with the GPT3 completions API but supposedly upgraded to the gpt-3.5-turbo API. But right now I'm just guessing that that is what happened.
Hi! I´ve started testing this extension and it is wonderful. However, I´d like to ask you if there is a way in which I could select with type of model (Ada, Babbage, Curie or Davinci) should my docs use, as the inquire of programming model capacity may not be always the same, making it more manegable and cost saving in cases where Ada could be used but Davinci is employed instead. For example, as a beta test of this extension I classified 750 names with a +GPT_Classify("Masculino o Femenino"), and this could have been done by Ada (According to OpenAI documentation which states that Ada is "Good at: Parsing text, simple classification, address correction, keywords"). As thy the size of the sample, the cost is not a problem, but given the chance of using it on larger data sets, being able of electing with model should GPT use would be great! Maybe this option is available, if so, could you tell me where? If not, I think this could be a great implementation to be done! Nevertheless, amazing work! It saved me a lot of time
Hello, you can absolutely choose the model, it is the last parameter of the functions. You can see our reference here: gptforwork.com/gpt-for-sheets/reference#6fc7eba8f45d4e87bb68d676e6893ea4
Hello! GPT for Work is not free but you can benefit from a free trial to test it before you buy a pack to add money to your balance. Find more information about our pricing here: gptforwork.com/help/billing