BasicallyOk
commented
4 months ago Need to implement ai.handle_submission(), which checks if any information required has not been touched on and flag submissions if this is the case. If the submitter later adds any information that is needed, the flag is removed (potential for an exploit there, but we're not going to implement submission review).
Is your feature request related to a problem? Please describe. We don't have categorization. I imagine it's something researchers would need.
Describe the solution you'd like Prompting the LLM to do it for us. We also need to prompt the LLM to ask questions if the submitters have yet to touch on those.
Additional context Erowid's categorizes submissions (based on submitter's input, which we are trying to streamline). Here's one.