Large language models (LLMs) have become central to modern artificial intelligence, yet their internal decision-making processes remain difficult to interpret. As interest grows in making these models more transparent and reliable, topological data analysis (TDA) has emerged as a promising mathematical approach for exploring their structure. This scoping review maps the current landscape of research where TDA tools—such as persistent homology and Mapper—are used to examine LLM components like attention patterns, latent representations, and training dynamics. By analyzing topological features across layers and tasks, these methods provide new ways to understand how language models generalize, respond to unfamiliar inputs, and shift under fine-tuning. The review also considers how TDA-based techniques contribute to broader goals in interpretability and robustness, especially in detecting hallucinations, out-of-distribution behavior, and representational collapse. Overall, the findings suggest that TDA offers a rigorous and versatile framework for studying LLMs, helping researchers uncover deeper patterns in how these models learn and reason.
The huge development of social networks opens the need for their analysis. The visualization of social networks is an important and powerful tool for such kind of analysis. There are many software packages which can be used for visualization. In this paper we are using graph theory, Python, and the free software Gephi to make visualization of social networks. Particularly we have made visualization of social network Twitter in order to show the relations between its users. This visualization helps in the analysis of relations and interactions between social media users.
27 January 2025 in International Journal of Science and Research (IJSR)
The newly emerged corona crisis in our country, but also much broader – on the entire planet, caused by the pandemic scale of COVID-19 virus, dictated the need for adjustment of the teaching and examination process of many university courses. At our institution, Faculty of Computer Science and Engineering (FCSE) in Skopje, starting from March 17, 2020, until today (March 2021), classes and exams are completely realized through distance learning systems, i.e. using the BigBlueButton video conferencing system, implemented in the Courses and Exams student services – the official FCSE websites on the Moodle e-learning platform. For all faculty courses, lectures, auditory and laboratory exercises, colloquia and exams, all take place via a video conferencing system for distance education. In this paper we present a comparative analysis of the conduction of some courses at FCSE in classical conditions, as opposed to the conditions with distance education. We have considered the analysis mainly from the aspect of the approach to teaching, as well as from the aspect of exam conduction and achieved exam results. The analysis of those aspects leads us to conclusions about several positive and negative sides that we noticed in distance education compared to the classical conditions of classes and exams. Our findings also may apply on the organization of online contests, especially in informatics.