11 · Practice: testing, bias & honesty

Fill in the missing word from Parts 2 and 3. Spelling doesn't have to be perfect — the explanation shows the idea.

1. 1. To really know if your model is good, you should test it on examples it has ___ seen before.

   Check

2. 2. The score for how often a model gets the right answer is called its ___.

   Check

3. 3. When a model memorizes its training examples but flops on new ones, that's called ___.

   Check

4. 4. When a model learns unfair patterns because its examples left people out, that unfairness is called ___.

   Check

5. 5. A good maker is ___ about what a model can do — saying 'right about 9 of 10 times,' not 'never wrong.'

   Check