Generative AI & LLMs
Bag of words
Describing a piece of writing by which words appear, without preserving their full order.
Example
A prediction system represents messages by counts of selected words.
Why people use it
It offers a simple way to compare texts by the words they contain.
What you'll hear
“Count the words without keeping their full order.”
What this means for you
Consider whether the task needs background information beyond word counts.
Can you control it?
Developer-only
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Can a word's frequency matter?
- Yes. Some versions count how often each word appears, rather than only recording whether it appears at all.
- Can it tell 'dog bites person' from 'person bites dog'?
- Not from word counts alone. Those sentences contain the same words but mean different things.
- Can common words overwhelm useful ones?
- They can, which is why systems often adjust which words count or how much weight they receive.