QCRI/AZERG-T4-Mistral
QCRI/AZERG-T4-Mistral is a 7 billion parameter language model fine-tuned from mistralai/Mistral-7B-Instruct-v0.3. Developed by QCRI, this specialist model excels at Task 4: Relationship Type Identification, specifically classifying STIX relationship types (e.g., uses, targets) between entities. It is optimized for high performance in entity detection within the AZERG framework, making it suitable for extracting STIX entities from security text passages.
Loading preview...
Model Overview
QCRI/AZERG-T4-Mistral is a specialized 7 billion parameter language model, fine-tuned from mistralai/Mistral-7B-Instruct-v0.3. Its primary function is Relationship Type Identification, specifically designed to classify STIX relationship types between two related entities within a given text.
Key Capabilities
- STIX Relationship Classification: Identifies specific STIX relationship labels (e.g.,
uses,targets,indicates) between source and target entities. - Cybersecurity Text Analysis: Optimized for processing security text passages to extract actionable threat intelligence.
- AZERG Framework Integration: Developed as part of the AZERG framework for automated STIX entity and relationship extraction.
Intended Use Cases
This model is ideal for developers and threat intelligence analysts who need to:
- Automatically extract potential STIX entities from security-related text.
- Classify the precise nature of relationships between identified entities in cybersecurity documents.
Training Details
The model was fine-tuned using the QCRI/AZERG-Dataset, which focuses on relationship type identification tasks relevant to cybersecurity threat intelligence.