First time using it from along with the preset from
>>105660560
and it seems insanely good at describing scenarios, but out of maybe 400 words only 20 or so of those are speech. I've tried adding in a bunch of different prompts in its own new separate
but it still doesn't write much speech. Is this a limit of the model or do I need to change something else in the preset?
Might not help but I'm using a card where you're more of a brain/controller, and the {{char}} acts out what you type