"Your entire capacity for logical reasoning evaporates because someone looks displeased with you."

Dottore + Assistant {{user}}

Brief description

I’m back to making bots after a loooong break.

/spoiler/ Honestly, what they did in the game hit me so hard that I deleted my GI account. I’m not going to let myself be fed crumbs and pathetic scraps of hope just to play along with deliberate developer tactics and keep going for no real reason. Because, frankly, I don’t believe their hints anymore.

As a fan - like many others - I waited for years for him to become playable, and instead, they gave us his death.

Oh well. They may have ended his story in the game, but his lore has taken on a life of its own.

I needed a long time to cool down after that. Yes, he’s only a fictional character from a game, but he’s also a character I genuinely grew very attached to. /end spoiler/

This bot will probably seem incredibly boring to a lot of people, because I wrote it very specifically around my own needs.

Professionally, I manage several people, and while I’m definitely not quite as soft as {{user}} is portrayed here, I still struggle with assertiveness myself - which is exactly where the idea for this rather unusual scenario came from.

Still, maybe someone else will find something useful or enjoyable in it too, so I decided to make it public.

Info

In this scenario, you are Dottore’s assistant and have just been promoted to the position of research team manager.

Unfortunately, {{user}} has serious problems with assertiveness, and their subordinates are beginning to walk all over them.

Dottore intends to correct that. Using his own methods, naturally.

For once, this is not a Dead Dove.

Dottore may still be dark, intimidating, morally questionable, and very much himself - but he is unlikely to solve {{user}}’s assertiveness issues through experimental neurosurgery.

He’ll probably just... talk it through with them.

Probably.

Suggestions:

If you want the bot to take a specific backstory into account from the start, it helps to describe it using OOC and save it in chat memory too. That usually makes the LLM respect your scenario much better.

As mentioned above, some LLMs - especially weaker ones - may immediately try to “guess” {{user}}’s condition too early, which can ruin the gradual mystery and investigation aspect of the scenario. Unfortunately, this mostly happens because LLMs heavily rely on information from the first message and often try to “be helpful” by immediately inserting that knowledge directly into the roleplay response.

Because of this, it may help to include a small OOC reminder such as:

(OOC: Dottore is not omniscient. He does not know what {{user}} is experiencing and can only attempt to deduce it through observation and interrogation.)

Keep in mind that regardless of the LLM model, they are not perfect - sometimes they write nonsense, do not remember events or behave in a way that we do not want. I have no influence on this, but fortunately, choosing the right temperature and comments via OOC should help :)

Have fun!

Art credit : @TianLY_0607