Gunnvant focused on improving documentation quality in the Nixtla/statsforecast repository, addressing a key issue with terminology used to describe point predictions in multiple model classes. By replacing the term 'mean' with 'fitted' throughout the docstrings, Gunnvant ensured that the documentation accurately reflected the model outputs, reducing potential confusion for users. This work, implemented using Python and Jupyter Notebook, did not alter any model behavior or APIs but enhanced the clarity and maintainability of the codebase. The update supports better user understanding and lays a foundation for clearer release notes, demonstrating attention to detail in technical communication and documentation.