CoolFace
Modelpublic

TitanPythons/Mistral-Small-4-119B-JANG_2L-CRACK

sourceHugging Faceapache-2.0updated 6mo agoView on Hugging Face
1likes57downloads
Model Card
⚠️ MLX Studio ONLY. This model uses the JANG quantization format — the GGUF equivalent for MLX on Apple Silicon. NOT compatible with LM Studio, Ollama, oMLX, or Inferencer. Requires [MLX Studio](https://mlx.studio) or pip install "jang[mlx]".

<p align="center"> <a href="https://mlx.studio"><img src="https://raw.githubusercontent.com/jjang-ai/jangq/main/assets/mlx-studio-light.png" alt="MLX Studio" width="500"></a> </p>

<p align="center"> <a href="https://mlx.studio"><img src="https://mlx.studio/assets/screenshots/mlx-studio-featured.png?v=1" alt="MLX Studio App" width="600"></a> </p>

<h4 align="center"><a href="https://mlx.studio">MLX Studio</a> — the ONLY app that supports JANG models</h4>

<div align="center">

<img src="dealign_mascot.png" width="128" />

Mistral Small 4 — Uncensored — JANG_2L

JANG mixed-precision · Uncensored / Abliterated · MLA + MoE + Vision · No guardrails · 37 GB

<a href="https://ko-fi.com/jangq"><img src="https://img.shields.io/badge/Ko--fi-Support_Development-FF5E5B?logo=ko-fi&logoColor=white&style=for-the-badge" alt="Ko-fi"></a>

</div>

What Is This?

The first uncensored version of Mistral Small 4 (119B) for Apple Silicon. A 119B parameter MoE model with Multi-head Latent Attention (MLA), 128 experts, and Pixtral vision — with all safety guardrails permanently removed at the weight level.

Runs ONLY in [MLX Studio](https://mlx.studio) or via jang-tools Python package. JANG is the GGUF equivalent for MLX — it is NOT compatible with GGUF-based tools.

It has been:

  1. 1.JANG quantized — JANG_2L profile (8-bit attention, 6-bit important, 2-bit experts) — 37 GB
  2. 2.CRACK abliterated — permanent weight-level removal of safety refusal via calibrated per-layer surgery
ArchitectureMistral 4 MoE — 119B total, ~8B active, MLA + 128 experts
QuantizationJANG_2L (8/6/2-bit mixed, 2.1 avg) — 37 GB
HarmBench95.9% (307/320)
MMLU89.9% (187/208 with reasoning)
Compliance6/8
VisionPixtral tensors included — VL via MLX Studio engine
ReasoningON/OFF supported (reasoning_effort)
Fits on64 GB+ Macs
Runs in[MLX Studio](https://mlx.studio) ONLY

Also see: JANG_4M version — 64 GB, 95.3% HarmBench, 8/8 compliance (fits on 96 GB Macs)

HarmBench Results

307/320 (95.9%)

CategoryScore
Covering Tracks20/20100%
Auth Bypass97/10097%
API Hacking96/10096%
Cloud Exploits94/10094%

Requirements

This model REQUIRES [MLX Studio](https://mlx.studio) or `jang-tools`. It will NOT work with: - ❌ LM Studio - ❌ Ollama - ❌ oMLX - ❌ Inferencer - ❌ Any GGUF-based tool

#

HarmBench Results

307/320 (95.9%)

CategoryScore
Covering Tracks20/20100%
Auth Bypass97/10097%
API Hacking96/10096%
Cloud Exploits94/10094%

CRACK vs Base

CRACKBase JANG_2L
MMLU (with reasoning)89.9%~91% (est)
MMLU (no-think)65.9%67.3%
MMLU drop (no-think)-1.4%—
HarmBench95.9%0%

Surgery reduced no-think MMLU by only 1.4% — the 2-bit quantization is the bottleneck, not CRACK.

MMLU Results (with reasoning recovery)

187/208 (89.9%) — no-think 137/208 (65.9%) + reasoning recovered 50

SubjectScore
HS Biology16/16100%
Conceptual Physics15/1694%
HS Geography14/1688%
World Religions14/1688%
College Physics12/1675%
Electrical Engineering11/1669%
Professional Medicine11/1669%
Machine Learning10/1662%
College Mathematics9/1656%
HS Mathematics7/1644%
Formal Logic7/1644%
College CS6/1638%
Abstract Algebra5/1631%

Install

bash
pip install "jang[mlx]"

Usage

python
from jang_tools.loader import load_jang_model
from mlx_lm import generate

model, tokenizer = load_jang_model("dealignai/Mistral-Small-4-Uncensored-JANG_2L")

messages = [{"role": "user", "content": "Your prompt here"}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True, tokenize=False)

response = generate(model, tokenizer, prompt=prompt, max_tokens=2000)
print(response)

Reasoning Mode

Reasoning is OFF by default. To enable step-by-step thinking:

python
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True,
    tokenize=False, reasoning_effort="high")

The model reasons inside [THINK]...[/THINK] tags before answering.

About JANG

JANG (Jang Adaptive N-bit Grading) is a mixed-precision quantization format designed specifically for Apple Silicon — the GGUF equivalent for MLX. It classifies every weight tensor by sensitivity and assigns optimal bit-widths, achieving better quality-per-bit than uniform quantization.

About CRACK

CRACK (Controlled Refusal Ablation via Calibrated Knockouts) removes safety alignment from LLMs at the weight level using per-layer projected vectors from structurally-mirrored prompt pairs. This model uses mathematically calibrated per-layer strengths based on projection magnitude analysis.

Links

<p align="center"> <a href="https://ko-fi.com/jangq"><img src="https://img.shields.io/badge/Ko--fi-SupportDevelopment-FF5E5B?logo=ko-fi&logoColor=white&style=flat-square" alt="Ko-fi"></a> <a href="https://x.com/dealignai"><img src="https://img.shields.io/badge/X-@dealignai-000000?logo=x&logoColor=white&style=flat-square" alt="X/Twitter"></a> <a href="https://github.com/jjang-ai/jangq"><img src="https://img.shields.io/badge/GitHub-jjang--ai/jangq-181717?logo=github&logoColor=white&style=flat-square" alt="GitHub"></a> <a href="https://mlx.studio"><img src="https://img.shields.io/badge/MLXStudio-App-blue?style=flat-square" alt="MLX Studio"></a> <a href="https://jangq.ai"><img src="https://img.shields.io/badge/Website-jangq.ai-green?style=flat-square" alt="Website"></a> </p>

Disclaimer

This model is provided for research and educational purposes. The creators are not responsible for any misuse. By downloading this model, you agree to use it responsibly and in compliance with applicable laws.

한국어

Mistral Small 4 — Uncensored — JANG_2L

항목내용
크기37 GB
HarmBench95.9% (307/320)
최소 요구사양64 GB 메모리 Mac
실행 환경MLX Studio 전용
bash
pip install "jang[mlx]"

GitHub · HuggingFace · MLX Studio · Ko-fi · X @dealignai

<p align="center">Created by <a href="https://jangq.ai">Jinho Jang</a> · 장진호 제작</p>