Instructions to use mlabonne/NeuralDaredevil-8B-abliterated with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlabonne/NeuralDaredevil-8B-abliterated with Transformers:

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

pipe = pipeline("text-generation", model="mlabonne/NeuralDaredevil-8B-abliterated")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mlabonne/NeuralDaredevil-8B-abliterated")
model = AutoModelForCausalLM.from_pretrained("mlabonne/NeuralDaredevil-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 mlabonne/NeuralDaredevil-8B-abliterated with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mlabonne/NeuralDaredevil-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": "mlabonne/NeuralDaredevil-8B-abliterated",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mlabonne/NeuralDaredevil-8B-abliterated

SGLang

How to use mlabonne/NeuralDaredevil-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 "mlabonne/NeuralDaredevil-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": "mlabonne/NeuralDaredevil-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 "mlabonne/NeuralDaredevil-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": "mlabonne/NeuralDaredevil-8B-abliterated",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mlabonne/NeuralDaredevil-8B-abliterated with Docker Model Runner:
```
docker model run hf.co/mlabonne/NeuralDaredevil-8B-abliterated
```

mlabonne

fatima113 commited on Jul 25, 2025

Commit

d005553

verified ·

1 Parent(s): 54fdf5a

add AIBOM (#13)

Browse files

- add AIBOM (8e271140778c6fea0930ee4c1d38b72c7f808373)

Co-authored-by: fatimaahmed <fatima113@users.noreply.huggingface.co>

Files changed (1) hide show

mlabonne_NeuralDaredevil-8B-abliterated.json +149 -0

mlabonne_NeuralDaredevil-8B-abliterated.json ADDED Viewed

	@@ -0,0 +1,149 @@

+{
+    "bomFormat": "CycloneDX",
+    "specVersion": "1.6",
+    "serialNumber": "urn:uuid:ec2a096a-6435-4b15-94a2-3cdfe51237d6",
+    "version": 1,
+    "metadata": {
+        "timestamp": "2025-07-10T08:46:54.205541+00:00",
+        "component": {
+            "type": "machine-learning-model",
+            "bom-ref": "mlabonne/NeuralDaredevil-8B-abliterated-137a94eb-5b88-59d1-88c3-66d6367d780d",
+            "name": "mlabonne/NeuralDaredevil-8B-abliterated",
+            "externalReferences": [
+                {
+                    "url": "https://huggingface.co/mlabonne/NeuralDaredevil-8B-abliterated",
+                    "type": "documentation"
+                }
+            ],
+            "modelCard": {
+                "modelParameters": {
+                    "task": "text-generation",
+                    "architectureFamily": "llama",
+                    "modelArchitecture": "LlamaForCausalLM",
+                    "datasets": [
+                        {
+                            "ref": "mlabonne/orpo-dpo-mix-40k-898bbffe-f59a-53fe-a5b7-846dc484fbab"
+                        }
+                    ]
+                },
+                "properties": [
+                    {
+                        "name": "library_name",
+                        "value": "transformers"
+                    }
+                ],
+                "quantitativeAnalysis": {
+                    "performanceMetrics": [
+                        {
+                            "slice": "dataset: ai2_arc, split: test, config: ARC-Challenge",
+                            "type": "acc_norm",
+                            "value": 69.28
+                        },
+                        {
+                            "slice": "dataset: hellaswag, split: validation",
+                            "type": "acc_norm",
+                            "value": 85.05
+                        },
+                        {
+                            "slice": "dataset: cais/mmlu, split: test, config: all",
+                            "type": "acc",
+                            "value": 69.1
+                        },
+                        {
+                            "slice": "dataset: truthful_qa, split: validation, config: multiple_choice",
+                            "type": "mc2",
+                            "value": 60
+                        },
+                        {
+                            "slice": "dataset: winogrande, split: validation, config: winogrande_xl",
+                            "type": "acc",
+                            "value": 78.69
+                        },
+                        {
+                            "slice": "dataset: gsm8k, split: test, config: main",
+                            "type": "acc",
+                            "value": 71.8
+                        }
+                    ]
+                }
+            },
+            "authors": [
+                {
+                    "name": "mlabonne"
+                }
+            ],
+            "licenses": [
+                {
+                    "license": {
+                        "name": "llama3"
+                    }
+                }
+            ],
+            "tags": [
+                "transformers",
+                "safetensors",
+                "llama",
+                "text-generation",
+                "dpo",
+                "conversational",
+                "dataset:mlabonne/orpo-dpo-mix-40k",
+                "license:llama3",
+                "model-index",
+                "autotrain_compatible",
+                "text-generation-inference",
+                "endpoints_compatible",
+                "region:us"
+            ]
+        }
+    },
+    "components": [
+        {
+            "type": "data",
+            "bom-ref": "mlabonne/orpo-dpo-mix-40k-898bbffe-f59a-53fe-a5b7-846dc484fbab",
+            "name": "mlabonne/orpo-dpo-mix-40k",
+            "data": [
+                {
+                    "type": "dataset",
+                    "bom-ref": "mlabonne/orpo-dpo-mix-40k-898bbffe-f59a-53fe-a5b7-846dc484fbab",
+                    "name": "mlabonne/orpo-dpo-mix-40k",
+                    "contents": {
+                        "url": "https://huggingface.co/datasets/mlabonne/orpo-dpo-mix-40k",
+                        "properties": [
+                            {
+                                "name": "task_categories",
+                                "value": "text-generation"
+                            },
+                            {
+                                "name": "language",
+                                "value": "en"
+                            },
+                            {
+                                "name": "pretty_name",
+                                "value": "ORPO-DPO-mix-40k"
+                            },
+                            {
+                                "name": "configs",
+                                "value": "Name of the dataset subset: default {\"split\": \"train\", \"path\": \"data/train-*\"}"
+                            },
+                            {
+                                "name": "license",
+                                "value": "apache-2.0"
+                            }
+                        ]
+                    },
+                    "governance": {
+                        "owners": [
+                            {
+                                "organization": {
+                                    "name": "mlabonne",
+                                    "url": "https://huggingface.co/mlabonne"
+                                }
+                            }
+                        ]
+                    },
+                    "description": "\n\t\n\t\t\n\t\tORPO-DPO-mix-40k v1.2\n\t\n\n\nThis dataset is designed for ORPO or DPO training.\nSee Fine-tune Llama 3 with ORPO for more information about how to use it.\nIt is a combination of the following high-quality DPO datasets:\n\nargilla/Capybara-Preferences: highly scored chosen answers >=5 (7,424 samples)argilla/distilabel-intel-orca-dpo-pairs: highly scored chosen answers >=9, not in GSM8K (2,299 samples)\nargilla/ultrafeedback-binarized-preferences-cleaned: highly scored chosen answers >=5 (22\u2026 See the full description on the dataset page: https://huggingface.co/datasets/mlabonne/orpo-dpo-mix-40k."
+                }
+            ]
+        }
+    ]
+}