Autoregressive
A language model that predicts the next word based on all the words that came before it, building text one word at a time.
In Plain English
Autoregressive models generate language by making predictions in sequence—like composing a sentence where each new word depends on everything written so far. The word "autoregressive" comes from statistics, but here it simply means the model looks backward at its own output to decide what comes next. This is how most text-generating AI works today: it sees 'The cat sat on the' and predicts 'mat' is likely next, then uses that prediction plus all previous words to guess the next one. It's a chain reaction where each link relies on all the previous ones.
💡Real-World Example
When you use your phone's text-prediction feature and it suggests the next word, it's using autoregressive logic. You type 'I'm going to the' and it guesses 'store' or 'beach' based on those first words. If you accept 'store,' the next prediction for what follows changes because now it considers 'I'm going to the store' as context.
Related Terms
What did you think of our explanation?