Earlier this year, we launched Nebula, a private, locally-hosted, open-source, large language model (LLM) platform. Since the launch we have been running around telling everybody about how nice Nebula is, and how much it can help you with research. Today, we can tell you that we practice what we preach, by using Nebula for a meta analysis of VU research studies on LLMs in the 2024-2025 period.

We used Nebula to categorize the studies on whether they were LLM-related or not, and on the type of LLM use. We found the Computer Science department as leader in terms of AI research output, with an overwhelming 47% (35 papers), followed by the Language, Literature, and Communication department with 20% (15 papers), and the School of Business and Economics with 9% (6 studies). See the complete infographic below. The public repository with the analysis can be found here.

Many thanks to Lara Baseggio and Radu Nicolae for carrying this project.