Hello everyone, gpt4 1106 has arrived, but the gpt 0314 provided better answers. Do you have the same experience?
Will the gpt4 0314 continue after June 2024?
Hello everyone, gpt4 1106 has arrived, but the gpt 0314 provided better answers. Do you have the same experience?
Will the gpt4 0314 continue after June 2024?
Hi and welcome to the Developer Forum!
Do you have an example of the difference using the same prompt?
Thank you for your response.
An example prompt alone may not be sufficient to determine which model is better; we might need some statistical analysis.
I often inquire about mathematical situations, and I’ve found that GPT-4 0314 performs better than 1106.
This is why I’m particularly concerned about the support after June 2024.
The model’s retirement has not changed since they first introduced the function-calling replacement in June and extended their initial “new model every month with actual production snapshots” you can still read about in the March API gpt-3.5-turbo blog (until they revise that also, just like even the quality 0301 can produce was “revised”).
It is better at performing tasks and taking programming and instructions, that is certain.
The quality and verbosity has changed to where the appearance of the output of newer models can be more satisfying for general use as long as your system prompt is no more than “you are ChatGPT”, but you take an AI pretrained on literature and scientific journals, and then fill its tuning with the common man’s thumbs up and lots of denials, the results are predictable.
Your response has inspired me. Perhaps these 0301, 0314, 0613 can continue to be served in the form of GPTs.