bloatmaxx

by meatrocket meatyrocket@proton.me discord: meat.rocket

| archive | presets | prompts |

---

for Anthropic, OpenAI, Google, Cohere and other LLMs.

---

---

presets

click here for todd's ST extension

---

// frequently asked questions //

---

wtf is this?

my RP presets that focus heavily on presentation & interactivity through HTML/markdown, which are also based on the nature of your roleplay
includes a SillyTavern extension to support HTML button inputs from AI (for kino text adventures)

includes toggles for the following:

• HTML panels
• Codeblock panels
• Interactive Text Adventure Buttons
• Additional writing prompts

why is it so fucking large? 500+ tokens is bloat enough

this originally included a bunch of fluff for multiple models to work with.
the instruction set alone for the HTML + options excluding any examples is about 2.5k tokens in total.
it's meant to be fully modular, so disable and enable whatever you need

how to install?

• download the desired preset below
• import the preset into your sillytavern's chat completion tab
• enable prompts/prefill as needed (assistant prefills are at the bottom of the preset)

what model do i use?

latest preset uses 3.5 sonnet v2 (2024-10-22), but supports 3.0 opus. 3.5 sonnet v1 has mixed consistency and results

how do you get images? how does it work?

enable the prompt and the corresponding prefill toggle (depending on your version of the preset)
it uses an external service ("pollinations.ai") to prompt an image generation via url

retarded meme preset

agreed

---

// downloads //

---

ST extension

Created by horsearmor

clickable html elements in ST, for HTML & text adventure prompts copy https://github.com/horse-armor/st-clickable-inputs into your Extensions manager

Disclaimer: This extension injects a small <instruction> prompt @ depth 0 May result in issues on blank cards or non-HTML based presets. If you don't need it, disable the extension manually.

---

---

claude presets

bloatmaxx 3.1 (25/10/24)

3.1: organized prompts to maximize instruction following while retaining consistency with dialogue and output length

i highly recommend the use of this extension, which squashes the User/Assistant exchanges into one user message. ultimately, i found that this format worked best for me. ymmv

i cannot remember the amount of changes i have made since my last prompt... ive made continuous changes it over the year and found that 3.5 sonnet v2's enhanced ability to follow instructions while roleplaying effectively now makes this worth publishing this preset. i would eventually like to hear feedback on how best prompting practices are nowadays

bloatmaxx panels generate 1-3 panels randomly, automatically preferring to generate less by design.

if you want to change the number of paragraphs in the output, try changing this line under "prose nudge"

the writing prompts & extras are all optional. edit them and disable/enable at your leisure

---

regex to remove <!--hidden--> tags:

/<!--.*?-->/ims

regex to remove <CoT> blocks:

/[\[\<]CoT[\>\]].*?[\[\<]\/CoT[\>\]]*/ims

regex to remove ALL <tags> only:

/<[^>]+>/g should keep other information contained outside of the tags but not within them

---

split QR presets

splitmaxx (27/06/24)

v3: revamped first CoT preset & second preset with a separate CoT prompt & bloatmaxx 2.0, this is my experimental take on the splitclover QR prompts, using gpt-4o or Claude 3.5 Sonnet & Claude 3.0 Opus in tandem creates a CoT with a faster model to output your kino responses with slower, better RP model (claude)

---

gpt-4 presets

note (25/10/24): if openai stopped make lobotomized models i'd probably have more presets here

omni 1.0 (19/05/24)

v1.0: normal v1.3, jailbroken specifically for GPT-4o. it works best with CoT (yes i hate the meme too but the model is "fast" so)

(4urbo) bloat v10 (31/03/24)

v10: updated html instructions, added a vision prompt and a brand new text adventure prompt originally created for GPT-4 Turbo

(just enable the corresponding prompt + prefill)

(works well with the ST extension)

---

other models

mistral-large preset

v1.6

---

credits to pixijb for inspiring the main prompt of normalmaxx. a4a for the anti-lewd prompt (though i got it from someone else) special thanks to anonaugus for the prefill suggestions

---

---

---

special thanks to:

• horsearmor (for the extension, proxy owner)
• mm (examples, providing local model access, proxy owner)
• branon (proxy owner, and for making this card)
• lurker (huge help with prompting)
• /aicg/ anons & jbmakies for suggestions + ideas
• sturdycord