Open
Description
That is, any fields configured in structured output that are not response, and shouldRespond. For example, in the PAIR summit experiment: https://deliberate-lab.appspot.com/#/e/955492ca-bc51-4558-b2a0-a278ec8c8d99
After every message, the mediator the rates shouldRespond categorical label. It'd be helpful to see these additional fields in the UI, especially when prototyping moderators.
Proposed solution:
Wrap all additional fields in an explanationResponse. Then, the pink text that experiments can see in the conversation are key value fields of all of the extra fields, e.g.
- explanation: asdfasdfasdf
- shouldRespond: yes
One problem with this proposed solution is that the experimenter won't be able to see fields that surface if shouldRespond = false..