New GPT-3 model: text-DaVinci-003
beta.openai.com
New GPT-3 model: text-DaVinci-003
1–10 of 216 posts
Re: New GPT-3 model: text-DaVinci-003
#2From their email sharing the news:
We're excited to announce the release of the newest addition to the GPT-3 model family: `text-davinci-003`. This model builds on top of our previous InstructGPT[1] models, and improves on a number of behaviors that we’ve heard are important to you as developers.
`text-davinci-003` includes the following improvements:
* It produces higher quality writing. This will help your applications deliver clearer, more engaging, and more compelling content.
* It can handle more complex instructions, meaning you can get even more creative with how you make use of its capabilities now.
* It’s better at longer form content generation, allowing you to take on tasks that would have previously been too difficult to achieve.
Re: New GPT-3 model: text-DaVinci-003
#3This one seems weirdly more verbose, in my brief testing.
Definitely uses more tokens on average than 002 when answering.
The writing style seems oddly-pretentious and yet somehow uninformed, perhaps there were too many copies of HN in the dataset?
I kid, I kid.
Re: New GPT-3 model: text-DaVinci-003
#4This one seems weirdly more verbose, in my brief testing. Definitely uses more tokens on average than 002 when answering. The writing style seems oddly-pretentious and yet somehow uninformed, perhaps there were too many copies of HN in the dataset? I kid, I kid.
have you tried tweaking the behavior, e.g. asking not to guess, to be brief, providing examples, constraining output with syntactic sugar, etc? e.g. it's like talking to a person, give feedback and you might get better result.
Re: New GPT-3 model: text-DaVinci-003
#5This one seems weirdly more verbose, in my brief testing. Definitely uses more tokens on average than 002 when answering. The writing style seems oddly-pretentious and yet somehow uninformed, perhaps there were too many copies of HN in the dataset? I kid, I kid.
Brilliant business move. Charge by the token, and then train the model to ramble! Pretty soon you'll have to buy it a beer before getting anything useful :)
Re: New GPT-3 model: text-DaVinci-003
#6This one seems weirdly more verbose, in my brief testing. Definitely uses more tokens on average than 002 when answering. The writing style seems oddly-pretentious and yet somehow uninformed, perhaps there were too many copies of HN in the dataset? I kid, I kid.
Just read the hot covid poll topic if you're looking for more good training data. ;)
I was really impressed by the 002 version. Looking forward to trying out 003 tonight!
Re: New GPT-3 model: text-DaVinci-003
#7Anyone know which applications it will improve other than better quality content writing?
Re: New GPT-3 model: text-DaVinci-003
#8Since it’s trained on more recent scrapes, does this mean that this new version will be self-aware? So if you ask it about “gpt-3” it’ll know what you’re talking about?
Re: New GPT-3 model: text-DaVinci-003
#9Since it’s trained on more recent scrapes, does this mean that this new version will be self-aware? So if you ask it about “gpt-3” it’ll know what you’re talking about?
"What is gpt-3?
GPT-3 (Generative Pre-trained Transformer 3) is an unsupervised, large-scale language model developed by OpenAI. It uses deep learning methods to produce human-like text, with the goal of automatically generating readable text. GPT-3 is capable of generating text from a prompt, completing tasks like question-answering, language translation, and summarization."
The question is me, the answer is gpt-3.
Re: New GPT-3 model: text-DaVinci-003
#10Since it’s trained on more recent scrapes, does this mean that this new version will be self-aware? So if you ask it about “gpt-3” it’ll know what you’re talking about?
It looks like only the model was updated, the training data only goes up to June 2021 (same as davinci 2 I think).
But GPT-3 came out in May 2020 so it should be “self aware” already.