I’ve been in the same boat for months, and haven’t found a solution.
I don’t know why “exposed_entities” is part of some default templates for assistant agents, but it doesn’t seem to work. I thought that “exposed_entities” would be maybe defined only when rendering a prompt template, but according to my ollama logs, that is not the case.
Im looking for how to find this now - I think this is the last piece in understanding how your context looks from a dashboard perspective.
I’m pretty sure ‘state data’ in context is just expand[exposed_entites], if only I had a way to do it myself and get length(of that expansion)… (Divide by three-flourish and it’s approximate token consumption)
Any luck in the last 7 months? I’m learning template syntax in an attempt to get to this. I can work with states, names and attributes, at least at a beginner level.
I can’t see how to find what is exposed and I can’t show aliases, which is the next thing I will want to do once I have the listed of exposed entities.
I’ve not found a solution to this, sorry. However, there is always more than one way to skin a cat. Pay @NathanCu’s Friday’s Party thread a visit and have him blow your mind with LLM related stuff for Home Assistant.