I wonder if setting a system message for gpt vision is possible?
I am using gpt4 to act as a teacher in a art installation. I would like to add the ability to have users sketch and have the ”teacher ” give feedback. Im using a highly specified system message for the text model now, but was wondering if i can switch to have gpt-4-1106-vision-preview do all the text AND the vision and also have my system message, or if i need to call on two different models for this + do some trickery to have regular gpt 4 get the image descriptions in text at least.
Any ideas, can we do system messages for gpt-4-1106-vision-preview yet?
Yes, you can use system prompt. I use one in mine. However, please note that
- GPT-4 Turbo with vision may behave slightly differently than GPT-4 Turbo, due to a system message we automatically insert into the conversation
- GPT-4 Turbo with vision is the same as the GPT-4 Turbo preview model and performs equally as well on text tasks but has vision capabilities added
- Vision is just one of many capabilities the model has
Also, it does not currently support function calling yet.
- Currently, GPT-4 Turbo with vision does not support the message.name parameter, functions/tools, response_format parameter, and we currently set a low max_tokens default which you can override.
See docs for more details
Interesting, I just tried adding a system message and it works. It was subtly changing it’s response in line with different system messages.