You were designated an "Abnormality" and captured by the Containment Agency.

(This bot is a WIP)

This bot is optimized so it DOES speak and act for your persona. It'll attempt to elaborate on your actions and utilize your dialogue and description in interesting ways, beyond what you instruct it to do.

It will also not bend to your persona's every whim, sometimes causing actions to fail or be countered by the world or other characters.

I recommend checking out my published chat "Prompt-Response Test" to see how it works. Using short and roundabout prompts works best to allow all characters to have volatility and agency.

An example for a prompt would be:

»»

YourName speaks for the first time since she was brought to this facility. She spoke in simple language, but she used language nonetheless, which could be understood by the researcher before her.

»»

Of course, you replace YourName with your persona's name.

From testing in my own chats, this works excellently, and most importantly:
- It was tested with JanitorLLM.
- Custom Prompt field is empty.
- Prefill is disabled.

This is sufficient, though if you want to curate or improve the behavior, using custom prompts and prefills might prove useful.

If you want to attempt to override the "speaks for you" part, you can try using custom prompts and prefills, but it'd be better to just ask me to make a version that does not force the AI to speak for you. Otherwise, leaving them empty seems to work perfectly for the purpose I have given it.

Now, in terms of Long Term Memory, here's a rentry explaining a lot of important things: https://rentry.co/jaimemtips

As a closing note, this thing's Limitless, as it should. If you want to drive a more lewd route, you can. NSFW should always be allowed. You can direct the LLM by using memory prompts or just by being really freaky on the chat. Things should resolve themselves in that direction if you do that.