Hello! I have been busy working on multiple plugins lately, and I came across a demonstration of the code interpreter and search model. I successfully recreated the search model and Python executor in a separate sandbox plugin, and after conducting some testing, I have some inquiries. Unfortunately, I do not have access to the other models, which prevents me from examining them myself.
- I am wondering if there is a way to establish an endpoint that executes XYZ and requests the user to input a file, similar to how the code interpreter operates. Is this currently possible?
- When choosing between Code Interpreter, Plugins, or GPT With Browsing, is there a distinction in the models? I understand that using Plugins involves a modified ada-002 model, but in terms of Code Interpreter and GPT With Browsing, are these models optimized for Python writing and generating more effective browsing queries?