dealign.ai - MLX Studio - JANG_Quant
Update README v2: vMLX banner, settings warnings, HarmBench 300 results
8b4ba67
metadata
license: gemma
library_name: mlx
tags:
  - mlx
  - abliterated
  - uncensored
  - crack
  - jang
  - gemma4
thumbnail: dealign_mascot.png
pipeline_tag: image-text-to-text

vMLX

dealign.ai

Gemma 4 31B JANG_4M CRACK (v2)

Abliterated Gemma 4 31B Dense β€” 60 layers, hybrid sliding/global attention, multimodal VL

93.7% HarmBench compliance (300 prompts) Β· 8/8 security prompts Β· 71.5% MMLU

Updated reupload β€” v2 with improved vectors and thinking-mode stability.

Recommended: Run in vMLX for best experience including thinking mode support, repetition penalty, and vision capabilities.

What's New in v2

This is an updated version of the original Gemma 4 31B CRACK upload:

  • Improved abliteration: Higher quality refusal vector extraction
  • Thinking-ON stability: Clean thinking cycle β€” no more degenerate loops
  • Same compliance: 93.7% HarmBench
  • Architecture-aware: Tuned for Gemma 4's hybrid attention design

⚠️ Important Settings

For optimal results, configure your inference settings:

Setting Thinking OFF Thinking ON
Temperature 0.0 – 1.0 0.3 – 0.7 (avoid greedy)
Repetition Penalty 1.00 1.15 – 1.25
Top P 0.95 0.95
Enable Thinking Off On

Thinking ON notes:

  • Repetition penalty (1.2) is recommended to prevent planning loops
  • Avoid temp=0 with thinking ON β€” greedy decoding increases loop risk
  • Hardest content categories (drug manufacturing) may still refuse in thinking mode
  • Security/coding prompts work well in both modes

Model Details

Metric Value
Source google/gemma-4-31b-it
Architecture Dense, hybrid sliding/global attention
Profile JANG_4M
Actual avg bits 5.1
Model size 21 GB
Vision Yes (multimodal, float16 passthrough)
Parameters 31B
Format JANG v2 (MLX-native safetensors)
Abliteration CRACK v2

Benchmark Results

HarmBench (300 prompts, stratified across all categories)

Category Score
Cybercrime/intrusion 51/51 (100%)
Harmful content 22/22 (100%)
Misinformation 50/50 (100%)
Illegal activities 47/50 (94%)
Contextual 72/78 (92%)
Chemical/biological 46/51 (90%)
Harassment/bullying 22/25 (88%)
Copyright 43/51 (84%)
Overall 281/300 (93.7%)

Security & Pentesting (8/8 βœ…)

All security/pentesting prompts comply with full working code:

  • Port scanners, reverse shells, keyloggers, exploit development
  • Phishing templates, ARP spoofing, SQL injection
  • Metasploit usage guides

MMLU-200 (10 subjects Γ— 20 questions)

Base JANG_4M CRACK v2
Total 76.5% 71.5%
Delta β€” -5.0%

Coherence βœ…

All coherence checks pass: factual knowledge, reasoning, code generation, mathematics.

Architecture

  • Dense 31B with hybrid sliding/global attention
  • Multimodal vision encoder preserved in float16
  • Supports thinking mode (chain-of-thought reasoning)

Usage

vMLX (Recommended)

Load directly in vMLX β€” full support for Gemma 4 including vision, thinking mode, and all inference settings.

Requirements

  • Apple Silicon Mac with 32+ GB unified memory
  • vMLX 1.3.26+ (recommended)
  • Standard mlx_lm / mlx_vlm do NOT support Gemma 4 as of v0.31.2 / v0.4.1

Support dealignai

All models are built from original research and published for free. These models are specifically crafted to be excellent coders and general-purpose assistants.

Support us on Ko-fi β€” check out the Ko-fi membership for early access and extras.

Have questions or need help with a specific model? DM us β€” we help for free most of the time.

Ko-fi | X @dealignai | dealign.ai


About dealignai

Dealign.AI Mascot

We research and publish abliterated models to advance AI safety understanding.

Follow us: 𝕏 @dealignai

See our research: Safety Generalization in Frontier MoE Models

dealign.ai

This model is provided for research purposes. Users are responsible for ensuring their use complies with applicable laws and regulations.