How does ChatGPT work?
ChatGPT (the product) is an assistant powered by a large language model (LLM) trained on large amounts of text. The model reads your input as tokens and generates a reply one token at a time, usually favoring statistically likely continuations.
That does not mean human-like understanding, and it does not guarantee factual accuracy — verify important information.
mej7.com is not an official OpenAI site. The lab explains the same ideas with simplified interactive visuals: tokenization, embeddings, attention, transformers, and generation.
Three intuitive steps
- Looks — splits and reads the text (tokens).
- Guesses — scores likely next tokens.
- Says — emits a token and repeats until done.