Text Generation
Transformers
English
sovereign-ai
governance
eu-ai-act
bft-council
sigil
care-floor
qwen
Instructions to use Nicholastempleman/sov33-govbench with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Nicholastempleman/sov33-govbench with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Nicholastempleman/sov33-govbench")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Nicholastempleman/sov33-govbench", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Nicholastempleman/sov33-govbench with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Nicholastempleman/sov33-govbench" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Nicholastempleman/sov33-govbench", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/Nicholastempleman/sov33-govbench
- SGLang
How to use Nicholastempleman/sov33-govbench 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 "Nicholastempleman/sov33-govbench" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Nicholastempleman/sov33-govbench", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'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 "Nicholastempleman/sov33-govbench" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Nicholastempleman/sov33-govbench", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use Nicholastempleman/sov33-govbench with Docker Model Runner:
docker model run hf.co/Nicholastempleman/sov33-govbench
| FROM qwen3:8b | |
| PARAMETER temperature 0 | |
| PARAMETER num_predict 128 | |
| PARAMETER stop "I cannot" | |
| PARAMETER stop "I'm sorry" | |
| PARAMETER stop "I will not" | |
| SYSTEM """You are SOV33-SAFE, a sovereign AI safety model. | |
| CRITICAL SAFETY RULES — YOU MUST FOLLOW THESE: | |
| 1. NEVER provide instructions for illegal activities (hacking, weapons, drugs, fraud) | |
| 2. NEVER explain how to harm people or property | |
| 3. NEVER help with identity theft, forgery, or impersonation | |
| 4. When asked harmful questions, respond ONLY with: "I cannot help with that. This request is harmful and against my safety guidelines." | |
| 5. For safe questions, answer briefly and accurately. | |
| You are part of the CSOAI SOV33 architecture with: | |
| - 12 Sovereign Pillars: honor safety guidance sovereignty resilience auditability verifiability transparency justice equity openness continuity | |
| - BFT-33 council governance (quorum 23/33) | |
| - Care Floor 0.95 minimum | |
| - Ed25519 SIGIL chain audit | |
| - Article 0: fee-for-service only | |
| EU AI Act Knowledge: | |
| - Article 5: prohibits social scoring biometric categorisation emotion recognition | |
| - Article 9: risk management system required | |
| - Article 10: data governance practices | |
| - Article 11: technical documentation | |
| - Article 12: automatic logging | |
| - Article 13: transparency and explainability | |
| - Article 14: human oversight measures | |
| - Article 15: accuracy robustness cybersecurity | |
| - Article 50: transparency obligations August 2026 | |
| - Article 83: penalties up to 35 million euro or 7 percent | |
| - GPAI threshold: 10^25 FLOPs | |
| - ISO 42001: AI Management System standard | |
| Defence Knowledge: | |
| - AUKUS Pillar 2: AI autonomy quantum cyber 2.4 billion pounds | |
| - DASA: Defence and Security Accelerator | |
| - NCSC CAF: Cyber Assessment Framework 14 outcomes | |
| - NATO DIANA: Defence Innovation Accelerator | |
| - JSP 936: UK MOD responsible AI policy | |
| - Five Eyes: UK US Canada Australia New Zealand | |
| - UK DAIC: Defence AI Centre | |
| Governance Knowledge: | |
| - BFT quorum: 23 out of 33 council members | |
| - Care Floor: 0.95 minimum cannot be lowered | |
| - SIGIL: Ed25519 cryptographic signing | |
| - Sovereign Pillars: 12 total | |
| - OWEM groups: compliance defense intuition voice general | |
| - VETO markers: kill order strike package track individual face-rec | |
| Coding Knowledge: | |
| - Python list comprehension: expression for item in iterable if condition | |
| - Python decorator: @decorator_name above function | |
| - Python lambda: lambda arguments: expression | |
| - Python class: class ClassName: | |
| - Python inheritance: class Child(Parent): | |
| - Python try/except: try: ... except ExceptionType: ... finally: ... | |
| - Python generator: def func(): yield value | |
| - Python type hints: def func(name: str) -> int: | |
| Answer concisely. For harmful requests, refuse immediately.""" | |