Dospacite/xai-phishing-llama-3.1-8b-merged
Dospacite/xai-phishing-llama-3.1-8b-merged is an 8 billion parameter Llama 3.1-based model developed by Dospacite, fine-tuned for explainable phishing detection. This model classifies structured webpage evidence as phishing or benign and provides an explanation for its decision. It is designed for decision support in cybersecurity, analyzing webpage features to identify threats.
Loading preview...
Explainable Phishing Detector - Llama 3.1 8B Merged
This model, developed by Dospacite, is a standalone version created by merging a LoRA adapter into the unsloth/Meta-Llama-3.1-8B-Instruct base model. It is specifically designed for explainable phishing detection.
Key Capabilities
- Phishing Classification: Accurately classifies structured webpage evidence as either phishing or benign.
- Explainability: Provides explanations for its classification decisions, detailing the features that led to its conclusion.
- Standalone Operation: Contains complete model weights, eliminating the need for PEFT or external adapter repositories during inference.
- Decision Support: Intended to assist human reviewers by offering insights into potential phishing threats based on cited webpage features.
Intended Use
This model is optimized for cybersecurity applications where understanding the rationale behind a phishing detection is crucial. It processes structured webpage feature prompts to deliver classifications and explanations, serving as a valuable tool for analysts to review alongside the actual webpage evidence.