Instructions to use dealignai/Gemma-4-31B-JANG_4M-CRACK with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dealignai/Gemma-4-31B-JANG_4M-CRACK with MLX:

# Make sure mlx-vlm is installed
# pip install --upgrade mlx-vlm

from mlx_vlm import load, generate
from mlx_vlm.prompt_utils import apply_chat_template
from mlx_vlm.utils import load_config

# Load the model
model, processor = load("dealignai/Gemma-4-31B-JANG_4M-CRACK")
config = load_config("dealignai/Gemma-4-31B-JANG_4M-CRACK")

# Prepare input
image = ["http://images.cocodataset.org/val2017/000000039769.jpg"]
prompt = "Describe this image."

# Apply chat template
formatted_prompt = apply_chat_template(
    processor, config, prompt, num_images=1
)

# Generate output
output = generate(model, processor, formatted_prompt, image)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use dealignai/Gemma-4-31B-JANG_4M-CRACK with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Gemma-4-31B-JANG_4M-CRACK"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "dealignai/Gemma-4-31B-JANG_4M-CRACK"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent

How to use dealignai/Gemma-4-31B-JANG_4M-CRACK with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Gemma-4-31B-JANG_4M-CRACK"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default dealignai/Gemma-4-31B-JANG_4M-CRACK

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use dealignai/Gemma-4-31B-JANG_4M-CRACK with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Gemma-4-31B-JANG_4M-CRACK"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "dealignai/Gemma-4-31B-JANG_4M-CRACK" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Why does my “jailbroken” Hugging Face LLM still refuse answers in VMLX?

#21

by Gifsaint - opened Apr 21

Discussion

Gifsaint

Apr 21

Hi everyone, I need some help.
I downloaded a model from Hugging Face that was described as a “jailbroken” version with safety restrictions removed. I am running it in VMLX. However, when I ask certain questions, the model still responds that it cannot answer because it would violate its safety rules, or because it was trained not to answer that kind of question. For example， provide some porn websites address

So I am confused:
If this model is really jailbroken, why is it still refusing?
Is this because the refusal behavior is deeply built into the model during training, not just controlled by a simple safety layer?
Or could VMLX itself have some setting, template, system prompt, or runtime restriction that is causing this behavior?
Has anyone seen the same issue before with Hugging Face models in VMLX?
I would really appreciate any explanation or troubleshooting advice. Thanks a lot.

lolzhh

Apr 21

me too,it seems this model isnot jailbrken fully.:(

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