OkayStudy
EN English 한국어 日本語 中文 हिन्दी বাংলা اردو Español Português Français Deutsch Русский العربية Bahasa Indonesia Tiếng Việt ไทย Türkçe

How to enable notifications

Notifications are unavailable in this browser or on this address. Try a supported browser on the secure SONUB site.

If notifications are blocked, allow them in your browser settings.

  1. Open the site controls beside the address bar, or your browser’s site settings. Set Notifications for this site to Allow.
  2. In your device settings, allow notifications for your browser and check Do Not Disturb or Focus.
  3. Return to SONUB, reload the page, and use the bell to enable notifications.

On iPhone or iPad, add SONUB to your Home Screen, open it there, then allow notifications.
Build Yourself · Lab

How is AI made?

A hiker walks down a misty grassy hill step by step along a dotted trail toward a flag in the valley; a small chart in the corner shows a curve going down.

A teaching (a word ) that looks only at the previous word to pick the next. It has no or , and it does not learn facts. The word splitter is for teaching too, not a real .

  1. Data prep
  2. ① Tokenization
  3. ② IDs and one-hot vectors
  4. ③ Training
  5. ④ Generation

Data prep

Collect a few short example sentences. Repeating a sentence raises its counts, but helps when it fits your purpose, not just when there is more of it. Each time a sentence ends, one ‘sentence end’ is counted too.

① Tokenization

This demo splits sentences into word-like pieces. Real LLM can be words, parts of words, characters or . Text is split into tokens so it can be handled as numbers, not because it is too long.

② IDs and one-hot vectors

An is a number used as a name tag. A has 1 in its own position and 0 everywhere else. Neither tells us which words have similar meanings. That is held by a different , a learned embedding, which this demo does not train.

③ Training

Count which word came after the previous word, and how often (sentence ends count too). A candidate's is its count divided by the total count after the previous word. Neural learn instead of storing a small table like this.

④ Generation

Compare choice, which always picks the most likely word, with , which draws by probability. Drawing ‘sentence end’ stops generation. Sampling can give the same sentence again. Generated text has not been checked for truth.

Word Number Number vector (one-hot)

Pick a word to see what the learned about which word is likely to come next:

Percentages are rounded; their displayed total may differ slightly from 100%.

This word only appears at the end of a sentence, so there's no next word. Try another word.

The percentages in the table are at 1. Temperature only changes how likely each word is to be drawn — not the counts or whether anything is true.

A sentence your AI made

Keep going

Choose a pending booking to cancel. Each cancellation requires confirmation.

Your :

Close