GPT

GPT (Generative Pre-trained Transformer ✕ The transformer is the type of neural network that powered modern language AI. The "T" in GPT comes from it. More in the glossary → ) is the family of large language models ✕ An LLM is an AI model trained on huge amounts of text to understand and generate language. It is the engine behind assistants like ChatGPT. More in the glossary → developed by OpenAI that sits behind ChatGPT. Its name sums up how it works: it is generative, because it produces new text; it is pre-trained, because it learns from enormous amounts of text before being used; and it uses the Transformer architecture, the basis of most of today’s language AI ✕ Artificial intelligence is a computer system’s ability to perform tasks we associate with human intelligence, such as understanding language, recognizing images or making decisions. More in the glossary → . In essence, it works by repeatedly predicting the most likely next unit of text (a token ✕ A token is the smallest unit of text a language model works with: usually a word or a piece of a word. Pricing and limits are measured in tokens. More in the glossary → ), stringing together coherent responses. Its different versions (GPT-3, GPT-4 and later) have grown steadily more capable. Although "GPT" has become almost a synonym for AI chatbot ✕ A chatbot is a program you converse with by text or voice. Modern ones use AI to understand and respond in natural language. More in the glossary → , it is just one of the families of models that exist; other companies offer alternatives. It also shares the typical limitations of these models, such as hallucinations ✕ A hallucination is when an AI presents made-up or incorrect information as true, in a confident tone. It is why you should verify its answers. More in the glossary → .