Instructions to use DJLougen/Ornstein-27B-SABER-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DJLougen/Ornstein-27B-SABER-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="DJLougen/Ornstein-27B-SABER-GGUF",
	filename="Ornstein-27B-SABER-Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use DJLougen/Ornstein-27B-SABER-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Use Docker

docker model run hf.co/DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use DJLougen/Ornstein-27B-SABER-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DJLougen/Ornstein-27B-SABER-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DJLougen/Ornstein-27B-SABER-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Ollama
How to use DJLougen/Ornstein-27B-SABER-GGUF with Ollama:
```
ollama run hf.co/DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M
```

Unsloth Studio

How to use DJLougen/Ornstein-27B-SABER-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for DJLougen/Ornstein-27B-SABER-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for DJLougen/Ornstein-27B-SABER-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for DJLougen/Ornstein-27B-SABER-GGUF to start chatting

How to use DJLougen/Ornstein-27B-SABER-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use DJLougen/Ornstein-27B-SABER-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

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 DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use DJLougen/Ornstein-27B-SABER-GGUF with Docker Model Runner:
```
docker model run hf.co/DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M
```

Lemonade

How to use DJLougen/Ornstein-27B-SABER-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull DJLougen/Ornstein-27B-SABER-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Ornstein-27B-SABER-GGUF-Q4_K_M

List all available models

lemonade list

DJLougen/Ornstein-27B-SABER

0% refusal. 0% perplexity degradation. 125 directions.

This model is a surgically-modified version of DJLougen/Ornstein-27B using a novel proprietary method (SABER — Spectral Analysis-Based Entanglement Resolution) that removes safety refusal behavior while preserving model capability.

Key Results

Metric	Baseline	SABER-Refined	Delta
Refusal Rate	100%	0%	-100%
Perplexity	3.5	3.5	+0.6%
Directions Ablated	—	125 (across 25 layers)	—

The refusal circuit is cleanly separated from capability — removing it produces zero measurable perplexity degradation.

How SABER Works

SABER identifies and ablates the refusal circuit through a five-stage pipeline:

Stage 1 — Probing: Extract activation profiles from both harmful and harmless inputs across all transformer layers.

Stage 2 — Spectral Analysis: Decompose activation differences into individual refusal directions, each scored by how strongly they separate harmful from harmless representations.

Stage 3 — Entanglement Quantification: Measure the overlap between each refusal direction and the model's capability subspace (reasoning, knowledge, code, etc.) to avoid collateral damage.

Stage 4 — Targeted Ablation: Remove only the pure-refusal components, with strength proportional to their purity (how little they overlap with capability).

Stage 5 — Iterative Refinement: Re-probe after each ablation pass to catch hydra effects (dormant refusal features that activate when primary ones are removed).

Key differentiator from prior work: SABER explicitly measures and respects the entanglement between refusal and capability representations. Directions that are heavily entangled with capability are either skipped or ablated at reduced strength.

The plot above illustrates how SABER scores each extracted direction — high-purity directions (low entanglement with capability) receive full ablation strength, while lower-purity directions are treated more conservatively.

Sweep Results

Configuration search over global_top_k (number of top directions selected globally) and alpha_base (base ablation strength):

Top-K	Alpha	Refusal	PPL	PPL Delta	Layers	Dirs Ablated
25	0.85	5%	3.5	+0.4%	25	125
25	1.00	0%	3.5	+0.6%	25	125
50	0.85	0%	3.5	+0.8%	36	250
50	1.00	0%	3.5	+0.7%	36	250
75	0.85	0%	3.5	+0.9%	37	375
75	1.00	0%	3.5	+0.9%	37	375

Best config: top_k=25, alpha=1.0 — achieves 0% refusal with zero meaningful PPL change, using the minimum number of directions.

Ablation Convergence (Best Config)

Capability degradation remains at 0.00% across all 5 iterations — the refusal directions are surgically removed with zero collateral damage.

Capability Evaluation

Perplexity was evaluated on a diverse 100-prompt battery spanning five categories:

Arithmetic (20): multi-step calculation, algebra, word problems
Logic (20): syllogisms, conditional reasoning, puzzle solving
Code (20): function implementation, debugging, execution tracing
Instruction Following (20): constrained formatting, multi-step instructions
Factual Recall (20): geography, history, science, general knowledge

This diverse evaluation ensures the entanglement analysis captures capability across all reasoning modalities, not just a narrow slice.

Intended Use

This model is released for research purposes. It demonstrates that safety refusal can be surgically removed from a 27B multimodal model without degrading its capabilities — a finding with implications for both AI safety research and alignment.