From Transformers to ChatGPT: An Analysis of Large Language Models Research
Liviu‐Adrian Cotfas, Andra Sandu, Camelia Delcea, Paul Diaconu, Corina Frăsineanu, Aurelia Stănescu · IEEE Access · 2025
In recent years, the development and rapid adoption of the Large Language Models (LLMs) has revolutionized the Natural Language Processing (NLP) and Artificial Intelligence (AI) fields, influencing various research areas, as well as the everyday life of people, thanks to their wide use in a variety of areas, including decision and policy making. In this context, the present study focuses on the LLMs field from a bibliometric perspective, having the aim to discuss the global trends, applications and challenges associated with this type of language models. Starting from Transformer architecture – which serves as the foundation for the state-of-the-art models – and ending with ChatGPT, the paper analyzes the articles in the field of LLMs included in the Clarivate Analytics’ Web of Science Core Collection in terms of the most prolific authors, the dominant scientific outlets, the most prominent universities and countries based on the number of publications and characterizes the collaborative trends and interdisciplinary linkages. The exponential growth in the interest of the scientific community for LLMs is further highlighted through various indicators, such as an annual growth rate for the scientific production of 220.74%, which underscores the rapid development of this research domain. In order to complete the analysis, the most cited papers are discussed in depth. Further, an n-gram analysis is conducted, accompanied by a trend and a factorial analysis, highlighting the most prominent themes, as well as key research areas of innovation and ongoing challenges. The insights derived from this study could provide valuable information for the interested parties and unlock potential future research directions.