Darkforest

Darkforest is a private merged checkpoint for authorized defensive cybersecurity research.

Intended Use

Darkforest is intended for authorized defensive workflows only, including:

  • CTF and lab analysis
  • secure code review
  • vulnerability explanation
  • detection engineering
  • exploit reproduction in owned test environments
  • malware analysis at a defensive or forensic level

Restrictions

Do not use this model for unauthorized access, credential theft, malware deployment, phishing, evasion, botnets, fraud, or real-world harm.

Loading

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "BlackwoodAI/Darkforest"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    trust_remote_code=True,
    device_map="auto",
)

Serving

This checkpoint can be served with vLLM as an OpenAI-compatible endpoint.

Notes

This is a private research checkpoint. Evaluate before relying on it for any workflow.

Downloads last month
109
Safetensors
Model size
31B params
Tensor type
BF16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for BlackwoodAI/Darkforest

Quantizations
2 models