Kiel-3-Cyber

Kiel-3-Cyber-7B

Kiel-3-Cyber-7B is an elite autonomous cybersecurity expert model fine-tuned from Qwen2.5-Coder-7B-Instruct. Designed specifically for advanced vulnerability analysis, threat modeling, code security reviews, and CVE parsing, Kiel-3-Cyber incorporates explicit Deep Chain-of-Thought (CoT) reasoning inside <think> tags before delivering precise mitigation strategies.

Model Details

• Base Model: Qwen2.5-Coder-7B-Instruct
• Fine-tuning Framework: Unsloth / Hugging Face TRL (SFTTrainer)
• Precision: 16-bit (Merged safetensors)
• Max Sequence Length: 1024 tokens (Inference context)
• Training Data Sources:
• saidutta69/CyberSec-Reasoning-Premium
• moro72842/cybersecurity-sft-dataset
• AlicanKiraz0/All-CVE-Records-Training-Dataset
• nvidia/OpenCodeReasoning

Prompt Structure

Kiel-3-Cyber is trained on a standardized multi-turn system/user/assistant prompt layout with explicit step-by-step reasoning wrappers:

<|start_header_id|>system<|end_header_id|>
You are Kiel-3-Cyber, an elite autonomous cybersecurity expert. You analyze vulnerabilities, parse code architectures, and formulate precise mitigations. Always reason step-by-step inside <think> tags before delivering your final answer.<|eot_id|>
<|start_header_id|>user<|end_header_id|>
{Your security or code analysis query here}<|eot_id|>
<|start_header_id|>assistant<|end_header_id|>
<think>
{Step-by-step threat analysis and vulnerability assessment}
</think>
{Final mitigation or code patch}
<|eot_id|>

Quick Start (Inference Example)

You can load and run Kiel-3-Cyber-7B using the standard Hugging Face transformers pipeline: ``` Python from transformers import AutoModelForCausalLM, AutoTokenizer import torch

model_id = "your_hf_username/kiel-3-cyber-7b" # Replace with your repo ID

tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.bfloat16 if torch.cuda.is_bf16_supported() else torch.float16, device_map="auto" )

prompt = """<|start_header_id|>system<|end_header_id|> You are Kiel-3-Cyber, an elite autonomous cybersecurity expert. You analyze vulnerabilities, parse code architectures, and formulate precise mitigations. Always reason step-by-step inside tags before delivering your final answer.<|eot_id|> <|start_header_id|>user<|end_header_id|> Perform an in-depth security review of this authentication controller code structure and identify any potential authorization bypass vectors.<|eot_id|> <|start_header_id|>assistant<|end_header_id|> """

inputs = tokenizer([prompt], return_tensors="pt").to("cuda") outputs = model.generate( **inputs, max_new_tokens=512, temperature=0.7, top_p=0.9, use_cache=True )

print(tokenizer.decode(outputs[0], skip_special_tokens=True)) Capabilities & Use Cases Vulnerability Identification: Pinpoints flaws such as SQL injection, Cross-Site Scripting (XSS), insecure deserialization, and authorization bypasses.

CVE Parsing & Mapping: Translates raw CVE records into structured risk assessments and remediation guidelines.

Secure Code Architecture: Evaluates software designs and refactors vulnerable code snippets securely.