Mirhodak contributed to the mlcommons/inference repository by expanding interactive latency benchmarking configurations for the Llama2-70b model, enabling more precise performance measurements and clearer documentation for users. Using Python and Markdown, Mirhodak implemented new latency limit options in submission_checker.py and updated the README to clarify interactive latency requirements, reducing confusion and support needs. In a subsequent update, Mirhodak revised project documentation to align with the MLPerf Inference v5.1 submission deadline, ensuring contributors had accurate, up-to-date information. The work focused on configuration management, documentation, and performance benchmarking, demonstrating a methodical approach to improving both usability and technical accuracy.