You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It helps the chat bot to have a screenshot of the current ui and viewport. That way it can "see" what you are looking at. I have found this helps. So it would be good to have the following features:
Every time you submit a message, there are two "hidden" images that are added to the prompt. One is the interface (which I already made in a branch), and the other is the viewport. For the viewport I think we can use the same code from download current viewport in the snapshots panel, but that code is broken right now so I wasn't sure how best to do it.
In the chatbot overall, it would be good to have it run a hidden prompt when the conversation first starts (either first open of the chat if the chat is empty, or if the user hits refresh). Basically, it should generate a user message that is "hello, here's my current context, I will ask you some follow up questions" along with the two screenshots. This is hidden from the user. Then the first output from the assistant will be the opening of the chat. That way, the chat starts from the context already, and the chatbot has already said "hello".
The text was updated successfully, but these errors were encountered:
It helps the chat bot to have a screenshot of the current ui and viewport. That way it can "see" what you are looking at. I have found this helps. So it would be good to have the following features:
Every time you submit a message, there are two "hidden" images that are added to the prompt. One is the interface (which I already made in a branch), and the other is the viewport. For the viewport I think we can use the same code from download current viewport in the snapshots panel, but that code is broken right now so I wasn't sure how best to do it.
In the chatbot overall, it would be good to have it run a hidden prompt when the conversation first starts (either first open of the chat if the chat is empty, or if the user hits refresh). Basically, it should generate a user message that is "hello, here's my current context, I will ask you some follow up questions" along with the two screenshots. This is hidden from the user. Then the first output from the assistant will be the opening of the chat. That way, the chat starts from the context already, and the chatbot has already said "hello".
The text was updated successfully, but these errors were encountered: