Bilic/Mistral-7B-LLM-Fraud-Detection
Bilic/Mistral-7B-LLM-Fraud-Detection is a 7 billion parameter language model fine-tuned from Mistral-7B-v0.1 by the BILIC TEAM OF AI ENGINEERS. It leverages Grouped-Query Attention and Sliding-Window Attention, with a context length of 4096 tokens. This model is specifically optimized for analyzing conversation transcripts to determine if they are fraudulent or legitimate, making it suitable for fraud detection applications.
Loading preview...
Overview
Bilic/Mistral-7B-LLM-Fraud-Detection is a 7 billion parameter Large Language Model (LLM) developed by the BILIC TEAM OF AI ENGINEERS. It is a fine-tuned version of the original Mistral-7B-v0.1 model, specifically adapted for fraud detection tasks. The model was trained using AutoTrain on a diverse set of synthetically generated fraudulent transcript datasets.
Key Capabilities
- Fraudulent Transcript Analysis: Specialized in evaluating conversation transcripts to identify indicators of fraud.
- Instruction Following: Designed to respond to specific instructions, particularly for classification tasks related to fraud detection.
- Efficient Architecture: Incorporates advanced architectural features like Grouped-Query Attention and Sliding-Window Attention for improved performance.
Good for
- Automated Fraud Detection: Ideal for systems requiring automated analysis of textual conversations to flag potential fraud.
- Financial Services: Applicable in banking, credit card companies, and insurance for monitoring communications.
- Customer Service Monitoring: Can be integrated into customer support platforms to detect suspicious interactions.