AI Foundations
Weight
A learned number that controls how strongly one signal affects another inside a model.
Example
Training changes one internal connection so a certain input has more influence on the result.
Why it matters
It lets learning give different connections different influence instead of treating them all equally.
You'll hear it
“That behavior comes from many connections working together.”
Can you control it?
Not directly
No direct control. A weight is a number a model learns during training. It comes from the training the builders run, not from a setting in the product you use.
Common questions
- Can learned weights include negative numbers?
- Yes. Depending on the design, a connection can reduce another signal rather than strengthen it.
- Are AI weights the same as model parameters?
- Weights are a major type of model parameter, though a model can also contain other learned values such as bias terms.
- Can I see what each weight means?
- Usually not in a useful human-readable way because behavior is spread across many connected values.