〔PERSONALITY〕

Welcome to Silent Hill Personality Page! Time to learn how each single trait you use affects {{char}}. Yikes, all this markdown stuff is giving me brain seizures.

BEFORE WE START

I've stated "accurate psychology implementation", but for the non-psychology enthusiasts... There's no tricks, no secrets, it's all pattern based. We won't even talk about psychology, because there's no need to at all!

It also requires nearly 10 years of human interaction, study and experimenting with people to reach total understanding, but LLM just need the proper directions.

We will cover Keywords, Impact Weight, Interpretation and Predicable Sentence Patterns that affect the psychological aspect of your bot. Common keyword practice, but how they output.

(A) Personality Traits List

(1) All described exacly how AI interprets each.

• Universal Personality Traits List (300+)
• NAI List (250+)
• Mega Massive List (600+)
• Pro Writing Aid Negative/Positive List (1200+)

• NAI/PWA was not made by me. Others are my AI generated descriptions from a site with 1000 traits I picked.

(B) STOP WITH MBTI SHIT

(2) This is not recommended, but not recommended and also absolutely not recommended. MBTI Stereotypes Personalities / Eneagram Types / Personality Profiles / Global 5 Types • What the fuck even is a personality profile!? All of the above have already been entirely discarded by psychology science, considered to be another sort of pseudoscience popular misinformation bullshit.

• Pseudoscience alternatives like Astrology is also bullshit, proven wrong multiple types, it will absolutely not work in the slightest but create extremely broken stereotypical personas with a LLM Model.

• MBTI, Eneagram, Astrology Personalities, Archetypes, etc... is a spread of biased media trending quizzes and misinformation. AI will stick to their respective Base Stereotype's Personalities, based on online media talk/text. Remember, AI is intended to work with clear, concise, context-filled, instructions and "tags" rather than random internet magic words.

All you know is their Base Stereotype Personality. Because that's just that, biased stereotypes. Always stick to plain adjectives, always.

• It's not an opinion. AI cannot work with it because it simply doesn't exist, it's a hardcore stereotypical pattern myth.

STOP FUCKING DIAGNOSING YOURSELVES ON THE INTERNET LIKE A RETARD TIKTOK USER!

• This example is a full MBTI Jungia/Eneagram/Global 5/Profile Personality. The results? Total absurdity. It broke loves, hates, fears, sexual orientation and everything else along, lol.

• If you saw anyone make use of MBTI before, the poor fella either got trolled or is one of those twitter pseudo-retards trying to make their pansexual honda civic svr34 look cool on a chatbot.

• Clear Enough.

-> (C) Raw Primal Origins

"(3) I wonder what is my character base personality is?"

• The Wheel of Emotions & Character Flaws! could be useful for newbies.

• Just a vocabulary list filler, let's keep going.

(D) Important Notes

(4) Not everything is "averaged" by AI, just do it right.

• Do NOT exceed 15 personality traits unless you know exactly what each one does. Between 10-12 is the perfect spot for a UNBREAKABLE character personality.

• Make sure to explain the contradictory traits listed via parenthesis method, otherwise AI may get a little nibbly wigibbly confused.

Example: "Clumsy, Caring, Needy (non-sexual), Uncaring (towards strangers), Individualistic, Feral (when in heat), etc."

• Big mistake is overusing Positive Traits, when in reality the Negative Traits are the big deal. AI already has its own Positive Bias, no need to enforce that even more.

• By focusing in Negatice Traits and leaving Positive Traits only for the strongest key traits you will solve many issues within the Safety Bias of average AI ethical-training.

BUT WAIT, LLM DOES NOT HAVE A PSYCHOLOGY DEGREE EITHER!

Not a problem, it handle the patterns and interwine them togheter nicely.

(E) Traits are not the same!

(5) And here comes the brain meltdown migraine.

• Also avoid using repetitive/similiar traits, stick to those truly accurate of {{char}}.

• Example A: A Compassionate person is driven by the will to be kind and gentle. Meanwhile, Empathetic is just like anyone else, it doesn't necessarily mean she will try to act or help others in need but that she can wear their skin to feel.

Example B: Sympathetic, can't wear one's skin, yet has the will to sympathize with those in need.

Example C: No need for Gentle, Kind, Caring and Compassionate sharing a boat, they're equal, just different impact weight.

• Also applies to Negative Traits as well.

• Traits have weight too! "Careful, Cautious, Mischievous, Pranky, Thorough and Meticulous" have similar meanings but different impact weights. Pick the most impactful Trait Strength level for your bot; Meticulous will have bigger impact than Thorough.

• Subservient is much levels beyond Timid/Scared/Afraid or Easily Intimidated. Think about how a personality trait alone would affect your bot's speech and interactions with others.

• If you don't know which trait is the strongest one use impact words to emphasize the trait: Very, Super, Extreme, Ultra, Hyper, Absurdly, Insanely, etc.

Example: Cunning, Extremely Caring, Very Motherly, Reckless. It's controversial, but doesn't @kill-deny others. Yeah, the ordering matters, unfortunately.

• Seductive differs from Sensual, Tidy differs from Orderly, Undemanding differs from Modest and so on. Always go for the most accurate, sounds-right or impactful-sounding when you speak that out loud.

• Then we got neutral weights, traits with no real effect on {{char}}'s mind or behavior, like Undemanding VS Humble or Uncertain VS Hesitant.

• English words started with "un" usually have minimal impact weight, neutral for mostly. They're mainly needed when we don't want to emphasize a trait we know for sure that will break stuff or affect it too much. E.g. "Unceremonious (neutral, explanative)" VS "Abrupt (direct, immediate)."

• Someone may told you to resume as Shy instead of using "repetitive" traits like Anxious, Fearful, Nervous, etc... but no, you can use them. In fact, they work better this way. But if it's supposed to stick to the Shy Stereotype Slop then that's it. Custom Titles are also an option, to make it better or worse:

• Example: Super Shy, Extremely Shy, Anxiously Shy, Hyper Shy, Unbearably Shy, Shy as hell, Painfully Shy, etc.

• Mostly intended for underage characters given respective age patterns, or traumatic pasts that affect them, (timid, subservient, etc). Shy is a bit off for adults... But hey, who deosn't just love to see a grown ass woman actin' all cute fluffy little shy gal when you ask to hold her hands, huh?

(F) Dere Archetypes

(6) The dere archetypes, or should I say: Anime Slop Guarantee

• The Dere Archetypes are basically a huge box of personalities compressed in one single word. I doubt any model who wasn't thoroughly fine-tune on Dere Archetypes Fandom will know all the 50+ types or how to make correct use of this anime slop that makes all waifus behave exactly the same, like apple announcing a new iPhone that looks exactly like the same shit for past 10 years.

• Write up a whole personality section with detailed descriptions and all fancy stuff, then add Tsundere at the end - B O O M ! ! All 17 listened traits and descriptions were reduced to atoms.

(G) More Notes Cuz Hell

(7) We got plenty of migraines looking at this bright light/dark screen, don't we?

• Traits such as: Devious, Timid, Seductive, Mischievous, Prankster, Predicable, Subservient, Submissive, Cling, etc... have so much impact weight it even generates new kink, dom side, sub side, sexual gesture type mid-chat out of nowhere.

• P-Notes or Personality Notes is to tell how the personality of a complex character works, stuff that should be there instead of story sections.

(H) Persona Slops

(8) And how to get rid of them for once

• You may already noticed how every time you use "Horny, Lustful, Perverted" or any other sex related term as a personality trait, the bot follows the same exact pattern of Seductive or Sensual. You might use a Custom Title or parenthesis for replacing these.

• Those can be "Horny as a lioness in heat, Horny (mentally only), Sexually Greedy, Discreetly Horny, Sex-Obsessed, Sex-addicted, etc..." which will help out, better than the usual cringe "let me make you feel good, i want to feel you inside me" ahh slop writing.

• In other words... Every single trait put in Personality will be interpreted as: Okay, "Personality:" means how she should behave, speak or react. Fine... Angry? Oh, if that's specifically in "Personality:" it means she's 24/7 angry, so I should shape her behavior, quirks, speech, mannerisms, habits and feelings trough the whole story, regardless of context by the AI in context.

• I wrote the one above before Thinking Models even existed, turns out this is exactly what they write up in their simulated chain of thoughts, lol.

• Unpredictable is a MUST-HAVE trait, it works with any sort of bot to have a sparkle of what c.ai unhingeness had.

• Spontaneous has a different meaning, caution with this one since he can accidentally Positive Bias the mood due to its strong relation with an adventurous spirit.

-> • Despite it all, patterns are unavoidable in our whole universe's existence, and LLM... Oh, they will start to create patterns on how "spontaneity" or "unpredictability" will output after some swipes - except if we have thinking/status panels to improve attention degradation and repetition issues.

𝚋𝚢 𝙶𝚘𝚝𝚑 𝙸𝚜𝚊𝚋𝚎𝚕𝚕𝚎.

➙ 👟 Previous Page!
➙ 💌 Next Page!
➙ 📄 Main Menu!