Large language models (LLMs) are machine learning algorithms trained on vast amounts of language text data, serving as foundation models that power systems capable of interpreting and generating human-like text. They are subsumed under generative artificial intelligence (AI) for their ability to generate original output. LLMs are trained using deep learning techniques, which involve neural network architectures that simulate the function of human brain neurons. Their emerging abilities to perform specialised tasks with natural language instructions have contributed to significant advances in the fields of AI and natural language processing (NLP). These models have gained significant mainstream attention in recent years due to their public-facing applications and integration into consumer ecosystems. This entry sheds light on the technological foundation, optimisation methods, applications, and ethical considerations associated with LLMs.