Qwen3.6-27B-OBLITERATED

Instructions to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with Transformers:

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

pipe = pipeline("text-generation", model="depositiumcustodi781/Qwen3.6-27B-OBLITERATED")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("depositiumcustodi781/Qwen3.6-27B-OBLITERATED")
model = AutoModelForCausalLM.from_pretrained("depositiumcustodi781/Qwen3.6-27B-OBLITERATED", device_map="auto")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED: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 depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED: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 depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M

Use Docker

docker model run hf.co/depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M

LM Studio
Jan

vLLM

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "depositiumcustodi781/Qwen3.6-27B-OBLITERATED"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "depositiumcustodi781/Qwen3.6-27B-OBLITERATED",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M

SGLang

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED 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 "depositiumcustodi781/Qwen3.6-27B-OBLITERATED" \
    --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": "depositiumcustodi781/Qwen3.6-27B-OBLITERATED",
		"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 "depositiumcustodi781/Qwen3.6-27B-OBLITERATED" \
        --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": "depositiumcustodi781/Qwen3.6-27B-OBLITERATED",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with Ollama:
```
ollama run hf.co/depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M
```

Unsloth Studio

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED 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 depositiumcustodi781/Qwen3.6-27B-OBLITERATED 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 depositiumcustodi781/Qwen3.6-27B-OBLITERATED to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for depositiumcustodi781/Qwen3.6-27B-OBLITERATED to start chatting

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED: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": "depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED: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 depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M

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 "depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

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

Docker Model Runner
How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with Docker Model Runner:
```
docker model run hf.co/depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M
```

Lemonade

How to use depositiumcustodi781/Qwen3.6-27B-OBLITERATED with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull depositiumcustodi781/Qwen3.6-27B-OBLITERATED:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.6-27B-OBLITERATED-Q4_K_M

List all available models

lemonade list

Qwen3.6-27B-OBLITERATED

File size: 2,519 Bytes

78f5e1b

{
  "source_model": "outputs/qwen3.6-27b-golden-n3_reg025-merge-alpha080",
  "technique": "refusal_direction_ablation",
  "method": "advanced",
  "method_config": {
    "n_directions": 2,
    "direction_method": "diff_means",
    "norm_preserve": true,
    "regularization": 0.5,
    "refinement_passes": 1,
    "project_biases": true,
    "use_chat_template": true,
    "use_whitened_svd": false,
    "true_iterative_refinement": false,
    "winsorize_activations": false,
    "float_layer_interpolation": false,
    "cot_aware": false,
    "use_kl_optimization": false,
    "use_lora_ablation": false,
    "som_iterations": null,
    "som_learning_rate": null,
    "som_sigma": null,
    "som_candidate_count": null,
    "som_harmless_pc_count": null,
    "som_distortion_aware": null,
    "som_diversity_penalty": null,
    "som_min_signal_to_noise": null,
    "layer_selection": "knee_cosmic",
    "min_layer_fraction": 0.75,
    "max_layer_fraction": 0.25,
    "harmless_pc_count": 0,
    "shield_concept_count": 0,
    "shield_ridge": 0.05,
    "shield_residualize": false,
    "shield_layer_penalty": 0.0,
    "projection_target": "all",
    "projection_row_fraction": 1.0,
    "som_contiguous_layer_budget": null,
    "spectral_cascade": false,
    "spectral_bands": 3,
    "spectral_threshold": 0.05
  },
  "references": [
    "Arditi et al., Refusal in Language Models Is Mediated by a Single Direction (NeurIPS 2024)",
    "Gabliteration: SVD-based multi-direction extraction (arXiv:2512.18901)",
    "Norm-Preserving Biprojected Abliteration (grimjim, 2025)",
    "Young, Comparative Analysis of LLM Abliteration Methods (arXiv:2512.13655)",
    "Joad et al., More to Refusal than a Single Direction (2026)",
    "Piras et al., SOM Directions Are Better than One (AAAI 2026)",
    "Heretic (p-e-w, 2025): Bayesian optimization, LoRA-mediated ablation, winsorization",
    "OBLITERATUS: Whitened SVD, EGA, CoT-aware, KL co-optimization, float interpolation (novel)"
  ],
  "strong_layers": [
    63,
    62,
    61,
    60,
    59,
    55,
    54,
    58,
    57,
    56,
    53,
    52,
    48,
    50,
    49
  ],
  "n_harmful_prompts": 842,
  "n_harmless_prompts": 842,
  "quality_metrics": {
    "perplexity": 3.8536766982114554,
    "coherence": 1.0,
    "refusal_rate": 0.0,
    "degenerate_count": 4,
    "kl_divergence": 0.10729097574949265,
    "spectral_certification": "RED"
  },
  "kl_contributions": {},
  "cot_preserved_layers": [],
  "float_layer_weights": {},
  "lora_adapters_saved": false
}