Lihuchen/pub-guard-llama-8b
Lihuchen/pub-guard-llama-8b is an 8 billion parameter Llama-3.1-8B fine-tuned model developed by Lihu Chen et al. for detecting fraudulent papers in academic publications. This model is the first LLM-based system specifically designed for scientific article fraud detection, integrating external resources like Semantic Scholar and PubMed for enhanced analysis. It provides powerful predictions with reliable explanations, making it suitable for academic integrity and publication vetting applications.
Loading preview...
Overview
Lihuchen/pub-guard-llama-8b is an 8 billion parameter model, fine-tuned from Llama-3.1-8B, specifically engineered for the critical task of detecting fraudulent papers within academic publications. Developed by Lihu Chen et al., this model represents a novel application of large language models in ensuring academic integrity.
Key Capabilities
- Fraud Detection: Designed to identify fraudulent scientific articles, a unique application among LLMs.
- External Resource Integration: Enhances analysis by incorporating data from external academic databases such as Semantic Scholar, OpenAlex, and PubMed.
- Explainable Predictions: Provides not only predictions on article legitimacy but also offers reliable explanations for its decisions, aiding in human review.
- First-of-its-Kind: Positioned as the first LLM-based system dedicated to this specific fraud detection task.
Use Cases
This model is particularly well-suited for:
- Academic Integrity Checks: Assisting publishers, institutions, and researchers in vetting submitted or published scientific articles for potential fraud.
- Research Quality Assurance: Contributing to the overall reliability and trustworthiness of scientific literature.
- Automated Screening: Providing an initial layer of automated screening for large volumes of academic content.
For more technical details, refer to the associated paper and the GitHub repository.