Large Language Model (LLM)
An AI trained on huge amounts of text that writes by predicting the next word, over and over.
In everyday terms
ChatGPT, Claude and Gemini are LLMs. They've read a large slice of the internet and books, and they continue text one piece at a time.
For professionals
A transformer trained on next-token prediction over trillions of tokens, then post-trained to follow instructions. Capabilities emerge from scale.
Think of it like…
Autocomplete on your phone, scaled up a millionfold.
You've already seen it
Every chatbot, AI writing assistant and "summarise this" button.
Myth vs reality
Myth: LLMs know whether what they say is true.
Reality: They generate likely text. Likely and true usually overlap, but not always.
Quick check
At its core, an LLM generates text by…
- Searching a database
- Predicting the next token repeatedly
- Copying a stored answer
- Asking a human
Show answer
Predicting the next token repeatedly: Next-token prediction is the engine.
Builds on
Neural Network · Generative AI