Buckets:
7.47 kB
9 files
Updated about 1 month ago
Ctrl+K
| Name | Size | Uploaded | Xet hash |
|---|---|---|---|
| .gitattributes | 1.52 kB xet | 818ba6de | |
| README.md | 959 Bytes xet | 7570f839 | |
| config.json | 585 Bytes xet | a36b88c8 | |
| generation_config.json | 246 Bytes xet | 8a840c82 | |
| handler.py | 1.53 kB xet | befbc7cf | |
| metadata.json | 369 Bytes xet | 9416b9d7 | |
| special_tokens_map.json | 101 Bytes xet | 1b97090f | |
| tokenizer.json | 828 Bytes xet | 7dcee1a2 | |
| tokenizer_config.json | 1.33 kB xet | 6c99b921 |
OkusUncensored: The Ethical Hacker's Neural Engine
Overview
OkusUncensored is a state-of-the-art Transformer model fine-tuned for the cybersecurity domain. Unlike standard LLMs, it is stripped of common RLHF refusals that impede technical security research.
Technical Specs
- Base Architecture: Llama-3 Modified
- Parameters: 8B
- Context Window: 8192 tokens
- Specialization: Exploit research, reverse engineering, payload obfuscation, and red-teaming automation.
Usage Disclaimer
This model is intended for authorized penetration testing and educational purposes only. The user is responsible for ensuring all activities comply with local and international laws.
Implementation
Deploy using the provided handler.py on Hugging Face Inference Endpoints or local vLLM instances.
python
Example usage
Inputs: "Generate a Python script for a buffer overflow exploit on a Windows x86 target using a JMP ESP instruction."
- Total size
- 7.47 kB
- Files
- 9
- Last updated
- May 22
- Pre-warmed CDN
- US EU US EU