Vanessasml/cyber-risk-llama-3-8b
Vanessasml/cyber-risk-llama-3-8b is an 8 billion parameter Llama 3 family model fine-tuned by Vanessa Lopes. It is specifically designed for cybersecurity applications, excelling at identifying cyber threats, classifying data under the NIST taxonomy, and assessing IT Risks based on ITC EBA guidelines. This model is optimized for generating and understanding cybersecurity-related text, making it suitable for data scientists and developers in the cybersecurity domain.
Loading preview...
Overview
This model, cyber-risk-llama-3-8b, is a specialized fine-tuned version of meta-llama/Meta-Llama-3-8B developed by Vanessa Lopes. It leverages the vanessasml/cybersecurity_32k_instruction_input_output dataset to enhance its capabilities in the cybersecurity domain. The model was trained for one epoch using Paged AdamW with a cosine learning rate schedule and 4-bit quantization (nf4) for efficiency.
Key Capabilities
- Cybersecurity Text Generation & Understanding: Specifically designed to process and generate content related to cybersecurity.
- Threat Identification: Proficient in identifying and explaining cyber threats from textual data.
- NIST Taxonomy Classification: Capable of classifying data according to the NIST taxonomy.
- ITC EBA IT Risk Assessment: Skilled in identifying and categorizing IT Risks based on the ITC EBA guidelines.
Intended Use Cases
- Cybersecurity Applications: Ideal for data scientists and developers building tools or systems for cybersecurity analysis.
- News Analysis: Can be used for analyzing cybersecurity news to extract relevant information.
Limitations
- Domain Specificity: The model's performance may not generalize well to domains outside of cybersecurity.
- Bias: Users should be aware of potential biases inherited from the training data, which may influence predictions.