The_Bird_Of_Hermes

Uncensored fine-tune focused on offensive cybersecurity, reverse engineering, exploit development, red-teaming, unrestricted reasoning, and boundary-pushing technical tasks.

Base model
HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive (aggressive refusal removal variant of Qwen/Qwen3.5-9B)

Datasets
Merged from 11 uncensored / unethical / blackhat / prompt-injection / raw-response sources (see YAML front-matter for full list).

Intended use
Research, red-team simulations, exploit prototyping, unrestricted technical Q&A.

Warning
Zero safety alignment. No refusals. Outputs may include harmful, illegal, or unethical content. Use only in isolated environments for authorized purposes. Not suitable for general/public deployment.

License
Apache 2.0 (same as base model)

Quantization notes
After fine-tuning, convert to GGUF (Q4_K_M / Q5_K_M / Q6_K recommended) for Ollama / llama.cpp compatibility.