đź’¬ Language AI: Words, Prompts & Transformers
Preview
Tap to open PDF
How does language AI work with words?
Language AI works with small pieces of text called tokens. It turns them into numbers, uses a transformer to connect clues across the text, and predicts a useful response. It does not understand language exactly as a person does.
1. Break text into tokens
A token can be a word, part of a word, or punctuation. A sentence becomes a sequence of these small pieces.
2. Turn tokens into numbers
Each token is represented by numbers so the model can compare it with patterns learned from many examples.
3. Connect the context
A transformer helps the model notice which earlier words matter for the meaning of the current word.
4. Predict what comes next
The model chooses likely next tokens one at a time to build an answer.
5. People check the meaning
A fluent answer can still be incorrect, so a person should check facts and whether the response fits the request.
Context changes meaning
The word bank is the same, but the nearby words help the model predict which meaning fits.
| # | Worksheet | Questions | Actions |
|---|---|---|---|
| Medium | |||
| 1 | 4 questions | ||
| 2 | 6 questions | ||
| 3 | 4 questions | ||
| 4 | 6 questions | ||
| 5 | 4 questions | ||
| 6 | Pro | 6 questions | |
Preview
Click to preview collection
Quick Tip
Teacher Resources
Differentiation
Learning Path
Skill Level
beginner
Estimated Time
20 minutes