Artificial Intelligence chatterbot and NLP (Natural Language Processing)

[Like Ra]

@shinybambi - doesn't Claude have any "explicit content" protections?

[shinybambi]

@shinybambi - doesn't Claude have any "explicit content" protections?

It absolutely does. But you can work around those protections when using API, providing it with system prompt with very explicit consent and a character card. It takes time to figure out proper system prompt, but its still possible.

Claude is absolutely obsessed with consents.

[spawn]

the llama3 is locked down pretty hard.
>>>BTW, do you use GGUF or PTH? Can't find an easy tutorial how to convert PTH to GGUF, and how to use multiple GGUFs for jan.ai. I'm learning... The good thing about jan.ai, is that it can use my NVIDIA card and it was easy to configure.

I go into hugging face and search for the model with gguf on the end of the search. I also noticed a lot of llama 3 model fine tuned to make it uncensored aeady.  The dolphin version should be pretty good.

Uncensored llama 3 models


I don't know anything about jan.ai.

[Like Ra]

The dolphin version should be pretty good.

Aha:

https://erichartford.com/uncensored-models

[brandynette]

ive skipped the 2015 posts & focused on 2023. WE GOT NOTHING!
https://adent.io/blog/best-ai-porn-generators/

nice to spend some free tokens to see what they can do!
as content creator id pic one fast to streamline creation & one detailed enough to get my kinks!

Well, as i see this from a junior fullstackers perspective we need a tiny ultra-fast LLM, pretrained no rails, ready to take in human language & train it on the content of ra´s forum!
once its trained on the forum we create variations of it with specific tasks in mind!

GOAL? FETISH LLM! its neurons need to talk perversion & depravity!

we need a decentralized, user shared resource stack to run an LLM literally just for us!
blockchain would come in really handy but we got no miners so no resources so not an option.

we need the ability to share workload among users no matter their hardware. where each image can allocate a certain number of neurons.
where every user can boot up the docker image on mouse click if they have docker installed.
a centralized machine coordinating the resources linked to the LLM without deciding for the LLM.

instead of trying to create the next fetish ChatGPT we make a tool for each task. tiny size, tiny prompts with as much fetish knowledge as posible!


1- chat bot: your prompt it with keywords & it spits out triggers. if you like it keeps pushing harder & deeper if you don't like it will get lighter & eventually request a new prompt.

2- image generator: only generates forum sized thumbs based on what it learned of the forum.

3- story like generator: it takes user details, identifies what fetishes link, overlap, hype each other & tries to provide a fappable 1000-word story.

10- Create user hypno content: instead of prompting with natural human language a drop-down list with specifically trained categories. like Categories, tags, meta tags, Sub tags & modifiers. this way we can train specific fetishes one at a time according to user wants instead of spending weeks on the whole at once!
for this we train it on our creators' content??? & we are doomed giggle!

7- it should be a tool creator like me can & want to use to streamline custom creation.
even if the LLM doesn't create the whole thing. it's still very useful providing with the parts of your creations.

like for my neon goth bambi.
https://brandynette.xxx/videos/neon-goth-bambi/

generated some neon goth bimbos on https://pornpen.ai/
generated some text based on my favorite https://www.reddit.com/r/BambiSleep/ post.
audio Input capture https://bambicloud.com/
put everything into OBS & hit record