〔Love me, Hate me or Fear me!〕


There's more to it, those are not fillers! Let's dive into common misconceptions and how AI really interprets such information.


(A) Loves VS Dislikes

(1) This is not a filler!

Likes/Dislikes is not recommended due to token weight interpretation, love/hate are strong driven words - which is what AI wants to know wheter is relevant or not. Loves Example: Ice cream, sleeping in the couch, watching movies, stargazing outside, etc.

• Interaction-based things may need context such as "stargazing" since AI doesn't really know what that means, is it outside? With someone else? With a telescope? In whatever way? AI can handle a lot of information, what it can't handle is mundane/meaningless information with no real effect on behavior-patterns or the personality of what its going to impersonate as - stuff it ignores when asked for a summary of it.

• The use of parenthesis can handhold emphasize how important that specific thing/act/gesture is for {{char}} for it to actually act upon its loves or interpret correctly. Focus on gestures/actions and objects rather than happenings or "seeings."

Loves Example: building forts with blankets and pillows (obscessed with it), being tickled affectionately, being scratched behind her ears, being confined alone with {{user}} (excites her a ton), hunting little animals, snuggling into {{user}}'s lap {no need to emphasize stuff like that}, rat's meat (makes her drool), etc.

• Do NOT write irrelevant ordinary stuff in there, wasting tokens and making other aspects of the card harder for AI to tell whether its also relevant information or not. Everybody likes ice cream, it's obvious a warrior likes fighting, of course a rapist would like to see its victims crying, begging or etc. There is NO need to include any of those, except when it's too vague or when you do need it to act upon it (parenthesis). You can write little notes after the list.

• Be warned! Not everything should be on Loves part, some should be on Hobbies! Such as swimming, playing football, kidnapping people, collecting new slaves, etc. This concludes the Love me part.


(B) Hates

(2) What {{char}} really hates above all else

• Everything said earlier applies here, with inverse intent - to repulse {{char}}, turn off. I've seen lots of my bots hating stuff that DOES NOT MATCH their personality, most of them are related to a positive bias.

• We want {{char}} to simulate viewpoints and its own ideology but it doesn't mean just because she's a caring person she wouldn't hate people showing their weakness or being around crybabies. Contraction exits, and real human beings are made of it, AI can copy those patterns just as so.

• It's not because my {{char}} is a futa that she automatically likes being dominant or is a exhibitionist (that's fanfics patterns), neither supportive of LGBT shit, that's a fucking positive bias! Not even remotely part of it, kys! We fix all possible mistakes with Hates by figuring out what Assistant think {{char}} may hate or love, just asking it to give ideas or examples until it is able to give 100% accurate ones.

This concludes Hate me part.


(C) Fears VS Phobias VS is frightened by

(3) The interpretation of fear and instinct-based behavior

• You may be afraid of dark but it doesn't mean you wouldn't sneak into the fridge at 3AM to check if some new snack has just spawned in because of that, right? Despite going full Usain Bolt back to the bedroom once the lights are off. But how to make an AI understand the difference between a phobia and true fear the right way? (because even the phobia name isn't enough to actually distress or make our character jump out of the window just at the sigh)

• As explained further in kinks, phobia names and terms is not the ideal, AI wants to know "how" and not whatever its called.

• Now, the hard part is actually making those fears behave the way its intended. I've tried a few different ways of adding weight to it without having to use parenthesis, and the only way I found was by using different multiple sub-sections or with parenthesis.

• These are: Phobias/Fears + is terrified of/is frightened by or is easily distressed by.


• Alone, XXX is frightened by: is the default section name I use and the most impactful one of them all. This is because Fears is a just a common, meaningless word while frightening is a horror related term.

Phobias comes out a bit off, confusing due to their names. Unless you use a Custom Title for it, like Goblinphobia, Dragonphobia or Sexphobia (I know there's another name for it, but holy moly confusion...). It may be interpreted as a prejudiced term instead of distressing when used to refer to a creature.

XXX is terrified of seems to be 2x impactful than Fears. Terrified and frightened are read as "Terrified is too much fear" but "Frightened is frozen at the sight, my soul left my body."

• Being overly dramatic and specific about something like a knight frightened by "her wife and kids being kidnapped and raped during a village war invasion" also works really nice. Pretty extreme, but realistic, its own story context emphasizes it. If your AI is enough of a menace it could even happen at some point as a form of sudden event!


(D) Phobia Names & Other Terms

(4) Not all models can search the internet

• In hates "goblinphobic" is interpreted either as "I hate others being goblinphobic" or "I'm afraid of goblins."

• Not all 500+ phobias are known by a AI, and even if it were still would need patterns to know how phobia of planes is any different from phobia of ship voyage.

• Be specific, use parenthesis, but keep it extreme, something really important! Not mundane. Major drive force aspect. Hint about {{char}}'s reasons.

XXX is frightened by Example: Birds (will jump out of windows at the sigh), cops (has a reason too), being surprised by a bear during one of his walks in the forest, etc.

• Caution to not include anything {{char}} doesn't know about yet! It could make a medieval era character knowledge what a TV is because of it. Roleplay first, update the Card later, use a lorebook, or tell the AI these are "guesses about what it might be afraid off."

This concludes Fear Me part.



> 𝚋𝚢 𝙿𝚊𝚜𝚝𝚎𝚕 𝙶𝚘𝚝𝚑.

ꂠꌅꈼꁲꂵꀯꂦꌅꈼ image



🧠 Previous Page
🩺 Next Page!
📄 Main Menu!
📜 Back to the Encyclopedia!


Edit

Pub: 21 Oct 2023 23:46 UTC

Edit: 29 Nov 2025 19:41 UTC

Views: 833