Domen Vake and dr. Aleksandar Tošić, InnoRenew CoE researchers, together with dr. Jernej Vičič and Bogdan Šinik from University of Primorska, recently published a scientific paper »Is Open Source the Future of AI? A Data-Driven Approach,« in Applied sciences.

Recently, large language models became central to academic research and industrial applications, fueling debates on their accuracy, usability, privacy, and potential misuse. Open-sourcing is often suggested as a means to enhance trust and transparency, yet, it comes with its own challenges, such as risks of illicit applications, limited financial incentives, and intellectual property concerns. Authors adopted a data-driven approach to examine the open-source development of large language models.

They assessed how community efforts impact model performance, and found out that the open-source community can significantly enhance models, demonstrating that community-driven modifications can yield efficiency gains without compromising performance.