Large Language Model (LLM)

Large language models are neural networks with billions of parameters that have been trained on vast corpora of text data from books, websites, code repositories, and other written sources. Through this training, they develop a statistical understanding of language patterns that allows them to generate coherent text, answer questions, translate languages, write code, and perform a wide range of language-related tasks.

The "large" in LLM refers both to the scale of training data and the number of parameters in the model. Frontier LLMs from OpenAI, Anthropic, and Google are trained on trillions of tokens of text. Their vendors no longer publish parameter counts for the largest closed models, so treat any specific number you read for them as an estimate. This scale enables emergent capabilities where the model can perform tasks it was never explicitly trained to do, such as reasoning through multi-step problems or following complex instructions.

LLMs work by predicting the most likely next token in a sequence, but this simple objective at massive scale produces surprisingly sophisticated behavior. They are the foundation of most modern AI applications including chatbots, coding assistants, content generators, and AI agents.