 What are n-grams? n is a number, g is tokens, like a word or a character, and model is probabilities. They were introduced way back in 1948 and provided a way to represent language with mathematics. A unigram model uses a single word probability to represent language. A bi-gram model uses two word probabilities to represent language, and so an n-gram model uses n-word probabilities to represent language. Although the use of n-gram models has gone down in recent years, they eventually paved way to the language models that we see today, so they are important to know about. Check out the channel for more videos on language models, and even building your own with code.