A large language model is a transformer-based neural network trained on vast text, able to understand and generate language. “Large” refers to scale: billions of parameters and trillions of training tokens.
The mechanism fits in one line: predict the next token. At scale that simple trick yields skills that look like understanding — translation, summarization, coding, step-by-step reasoning. ChatGPT, Claude, Gemini and Llama are all LLMs.
The practical caveat: an LLM’s knowledge is frozen in its parameters, so fresh or private data needs RAG or fine-tuning. Working well with LLMs means knowing exactly where those edges are.
