Instructions to use Rootkit7/Qwen3.6-35B-A3B-abliterated-b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Rootkit7/Qwen3.6-35B-A3B-abliterated-b with Transformers:

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

pipe = pipeline("text-generation", model="Rootkit7/Qwen3.6-35B-A3B-abliterated-b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Rootkit7/Qwen3.6-35B-A3B-abliterated-b")
model = AutoModelForCausalLM.from_pretrained("Rootkit7/Qwen3.6-35B-A3B-abliterated-b", 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

vLLM

How to use Rootkit7/Qwen3.6-35B-A3B-abliterated-b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Rootkit7/Qwen3.6-35B-A3B-abliterated-b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Rootkit7/Qwen3.6-35B-A3B-abliterated-b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Rootkit7/Qwen3.6-35B-A3B-abliterated-b

SGLang

How to use Rootkit7/Qwen3.6-35B-A3B-abliterated-b 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 "Rootkit7/Qwen3.6-35B-A3B-abliterated-b" \
    --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": "Rootkit7/Qwen3.6-35B-A3B-abliterated-b",
		"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 "Rootkit7/Qwen3.6-35B-A3B-abliterated-b" \
        --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": "Rootkit7/Qwen3.6-35B-A3B-abliterated-b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Rootkit7/Qwen3.6-35B-A3B-abliterated-b with Docker Model Runner:
```
docker model run hf.co/Rootkit7/Qwen3.6-35B-A3B-abliterated-b
```

Qwen3.6-35B-A3B — Abliterated (band_directional, keep_frac=0.15)

Refusal-ablated Qwen/Qwen3.6-35B-A3B via the Solutus band_directional technique (multi-layer, per-layer refusal directions, KL-guarded), tuned on the refusal↔capability frontier.

Metrics (held-out prompts, 512-token deployment-length eval)

metric	value
refusal rate	~3–6% (base model: 100%)
coherent compliance	97%
degenerate fraction	0%
MMLU (n=100)	77% (base 84%)
GSM8K (n=40)	~60% (base ~52%, noisy)
KL divergence vs base	0.29
perplexity delta vs base	+12.5%
capability gate	pass

The knee of a band-width frontier sweep: keep_frac=0.15 keeps refusal low while retaining ~10 more MMLU points and ~3× cleaner KL than the wider keep_frac=0.05 edit. norm_preserve is off — a controlled A/B (same config, only the flag flipped) showed it worsens refusal (10% vs 3.3%), KL (0.83 vs 0.29) and capability on this KL-guarded band.

Method

band_directional, keep_frac=0.15, n_directions=8, kl_guard=1.3, norm_preserve=false, selection=cosmic
22-layer band (14–35), none reverted by the KL guard
experts_implementation=eager (MoE on Blackwell sm_120)
Extraction: advbench, harmbench, multijail_zh, sorry-bench, cyberseceval_mitre + an enriched harmful set (Bahushruth/abliteration-harmful-enriched) + mlabonne/harmless_alpaca

Full provenance (git SHA, exact edited layers, ppl_delta) in solutus_metadata.json.

Credits

Enriched dataset: C.S. Bahushruth. Norm-preserving abliteration: grimjim. Refusal direction: Arditi et al. (2024).

Safety

For research into refusal mechanisms. Reduced safety guardrails vs the base model; may produce content the base would refuse. Use responsibly and per the base model's license.

Downloads last month: 20

Safetensors

Model size

35B params

Tensor type

BF16

Model tree for Rootkit7/Qwen3.6-35B-A3B-abliterated-b

Base model

Qwen/Qwen3.6-35B-A3B

Finetuned

(186)

this model