RefCheckAI

AI Reference Checker

AI Reference Checker automates citation validation for AI-generated content and academic research. It uses a custom model trained on a reference dataset to compare citation statements to source texts. This tool accelerates verification, improves accuracy, and provides contextual snippets for deeper understanding, transforming manual processes into an efficient automated workflow.

Key Features

• Classifies citations into four categories: supported, partially supported, not supported, and uncertain
• Provides semantic claim verification and extracts supporting evidence
• Reduces citation checking time from hours to minutes
• Improves research accuracy, quality, and reproducibility
• Easy-to-use web app interface

Resources

• Training Data
• Fine-tuned Llama-3.1-8B Model

PDF Docs

Team

The AI Reference Checker team consists of researchers and developers eager to improving the quality and efficiency of academic research.

Acknowledgements

We would like to thank the National Computational Infrastructure (NCI) for hosting the Generative AI CodeFest Australia together with NVIDIA, OpenACC Organization and Sustainable Metal Cloud (SMC), and a special thanks to our two expert mentors during this event:

• Hariharan Suresh (NVIDIA)
• Zhuochen Wu (NCI)

This program has been supported by the Australian Government Department of Industry, Science and Resources through the National AI Centre.

The project is supported by the Sydney Informatics Hub at the University of Sydney.

Contact

Early adopters and interested contributors, please email sebastian.haan@sydney.edu.au

Data and Model License

Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0), free for research and non-commercial use with attribution.