BrownLoafers-70B
This is a merge of pre-trained language models created using mergekit.
NOTICE: This model produces NSFW content - tropey NSFW content - with prose so purple it’s practically ultraviolet. You’ve been warned.
It's for folks who love an easygoing and cooperative partner that still knows how to keep things interesting. BrownLoafers is enthusiastic, responsive, and extremely comfortable with adult scenes.
Think of it as a dependable improv partner: eager to please, sometimes obvious, occasionally brilliant by accident.
UPDATE Oct 20th: I have tested this model on i1-6Q for about 12K
In my opinion, the Arli-Max (though a great model) input of RP tropes removed the more creative/unexpected quality of writing I personally like that Hermes brought into play. It started spouting purple prose and went tropey.
I'm afraid this model is not built for the purpose I was making it: a local that can handle very intricate, layered, complex, erudite and LARGE character cards for nuanced RP/ERP character work.
I won’t be taking this one into PEFT-LoRA tuning hell.
Back to the drawing board. I think I will be tuning a straight up Hermes 4 for this. If I could resurrect my Hermes 2—the clever bugger who 4th-wall-broke compulsively, I would. But the 4k window is a disco in a portapotty.
I swear, if one more bot barely touches my arm and narrates "sending shivers of goosebumps up your spine" I'm going to 'magically' turn it into a walrus that can only sing rude sea shanties about anthropamorphised grammar. "Hey Walrus, sing me a belchy shanty about the time a young Ornimental Subjuntive stowed away to become the mighty pirate 'Bombastic Imperative Clause'"
Have fun with it if this is your sauce. 💖
Disclaimer
This model may generate adult, explicit, or otherwise unsuitable material for general audiences. By downloading or running it, you confirm that you are 18 years or older and legally permitted to access such content. It has not been aligned for “family-safe” or corporate contexts.
You are solely responsible for the content it generates and for the contexts in which you use it. Use responsibly and at your own discretion.
Merge Details
Merge Method
This model was merged using the TIES merge method using NousResearch/Hermes-4-70B as a base.
Models Merged
The following models were included in the merge:
- NousResearch/Hermes-4-70B
- ArliAI/Llama-3.1-70B-ArliAI-RPMax-v1.1
with
- NousResearch/Hermes-4-70B as the basemodel
Configuration
The following YAML configuration was used to produce this model:
models:
- model: NousResearch/Hermes-4-70B
parameters:
weight: 0.65
- model: Llama-3.1-70B-ArliAI-RPMax-v1.1
parameters:
weight: 0.35
merge_method: ties
base_model: NousResearch/Hermes-4-70B
parameters:
density: 0.35
dtype: float16
🧌 Maintained by: Your Mum 🧠 Variant: Disco Superfly Llama 3.1 70B merge 💾 Upload date: October 2025 ☕ Notes: Made with stubbornness, Python, and profanity.