Dustflames

Night owl, daydreamer and prompt engineering hobbyist. Long time player of D&D, so my work includes RP elements.

Some of my stuff are direct import of my earlier experimentation from JAI, my shit is better for SillyTavern.

But if you're curious, here is my collection on [Janitor page] (https://janitorai.com/profiles/24826e3f-3458-4d06-a8cc-f0fec112ff02_profile-of-dustflames). I don't recommend it because I cut tokens and details for my bots there due to the site's limitations.

FAQ
What is an LLM? Large Language Model like Claude, OpenAI, Mistral, Cohere, etc. All of them behave differently because of the dataset they're trained in.
Which LLM do you recommend? I personally use Claude and LLama, but you should experiment to find what you like.
The bot is talking for me/repeating/has bad memory/too horny? That's an LLM issue. You can try things like adding prompts against it, and I do try to write the initial message that minimize the issues of speaking for users, but I don't develop and tune your LLM.
How come you don't have (insert whatever parts) in your character definition? Because sometimes less is more. Too much unnecessary details will confuse the poor little LLM. 12 inch cock with green leaf-like veins and horse like hairy balls and bubblegum head isn't exactly going to be implemented by all LLMs