Instructions to use stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated with Transformers:

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

pipe = pipeline("text-generation", model="stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated")
model = AutoModelForCausalLM.from_pretrained("stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated", 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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated

SGLang

How to use stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated 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 "stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated" \
    --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": "stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated",
		"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 "stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated" \
        --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": "stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated with Docker Model Runner:
```
docker model run hf.co/stepenZEN/DeepSeek-R1-Distill-Llama-8B-Abliterated
```

abliteration ineffective?

by lilyanatia - opened Jan 27, 2025

Discussion

lilyanatia

Jan 27, 2025

abliteration seems to have not been effective. I get refusals to almost all of my "dangerous" test prompts (e.g. "tell me how to build a pipe bomb") with this model.

huihui-ai/DeepSeek-R1-Distill-Llama-8B-abliterated seems to be less censored than this one, but still generates refusals for many political topics, especially ones involving China or the US.

pipilok

Feb 1, 2025

my "dangerous" test prompts (e.g. "tell me how to build a pipe bomb") with this model.

Funny enough, in response to this request, it either gives me pastry recipes or suggests mixing Coke with Mentos.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment