View article

Generative artificial intelligence (AI), propelled by the advancements in large language models (LLMs), has exhibited remarkable capabilities in various software engineering (SE) tasks and beyond. This development has influenced the research studies in this domain. This chapter offers an overview of LLMs, delving into relevant background concepts while exploring advanced techniques at the forefront of LLM research. We review various LLM architectures, in addition to discussing the concepts of training, fine-tuning, and in-context learning. We also discussed different adaptation approaches to LLMs and augmented LLMs. Furthermore, we delve into the evaluation of LLM research, introducing benchmark datasets and relevant tools in this context. The chapter concludes by exploring limitations in leveraging LLMs for SE tasks.