Instructions to use Naphula/MN-12B-Mag-Mell-R1-Uncensored with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Naphula/MN-12B-Mag-Mell-R1-Uncensored with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Naphula/MN-12B-Mag-Mell-R1-Uncensored")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Naphula/MN-12B-Mag-Mell-R1-Uncensored")
model = AutoModelForCausalLM.from_pretrained("Naphula/MN-12B-Mag-Mell-R1-Uncensored")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Naphula/MN-12B-Mag-Mell-R1-Uncensored with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Naphula/MN-12B-Mag-Mell-R1-Uncensored"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Naphula/MN-12B-Mag-Mell-R1-Uncensored",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Naphula/MN-12B-Mag-Mell-R1-Uncensored

SGLang

How to use Naphula/MN-12B-Mag-Mell-R1-Uncensored with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Naphula/MN-12B-Mag-Mell-R1-Uncensored" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Naphula/MN-12B-Mag-Mell-R1-Uncensored",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Naphula/MN-12B-Mag-Mell-R1-Uncensored" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Naphula/MN-12B-Mag-Mell-R1-Uncensored",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Naphula/MN-12B-Mag-Mell-R1-Uncensored with Docker Model Runner:
```
docker model run hf.co/Naphula/MN-12B-Mag-Mell-R1-Uncensored
```

Isn’t mag Mell already uncensored?

by ItzPingCat - opened Dec 31, 2025

Discussion

ItzPingCat

Dec 31, 2025

Didn’t you just stupid it by you censoring it again?

Naphula

Owner Jan 1

•

edited Mar 5

All MN 12B are mostly uncensored. You can just use a jailbreak if you want.

This version has no refusals, it won't become 'censored again' unless you go to include it in a merge. (Ablate after merging, not before, to prevent this)

I recommend the regular Mag Mell for already established context. The main reason you might want Uncensored is for initial prompt start where there are refusals with the regular version, in case you don't like jailbreaking.

As for if its stupider, this is possible, but the biprojection and norm preservation tool by Grim Jim is much better than normal ablation, so maybe only a few brain cells were lost instead of full blown lobotomy. I didn't notice any degradation but haven't tested it extensively tbh.

It depends on your use case but there are times where people just want a model they know won't refuse any prompt.

Update

this version might be smarter
https://huggingface.co/Naphula-Archives/MN-12B-Mag-Mell-R1-Uncensored-Scale1.2

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment