LLM Evaluation framework that leverages LLMs to Evaluate other LLMs
InferIQ is a NodeJS application designed to generate answers to questions using your sample dataset. After answering all dataset questions across the selected LLMs in your evaluation pool, a group of Judge LLMs will assess and rate each response. The results are then visualized in several graphs, providing a clearer picture of the overall accuracy of the LLMs. Additionally, metrics such as BERT Score and Inference Time are provided to further enhance your understanding of each LLM's performance.
- run the npm setup script:
npm run setup
- start the application:
npm run serverornpm start
Please consider sending me a tip to support my work 😀