Fix vectorized CountFilter

[SirYwell]

Overview

Description

The applyVector implementation of CountFilter just counted the number of lanes it sees, which doesn't match the behavior of the scalar implementation. By passing around a mask for the affected lanes, we can just count those lanes and get the same counts as without vectorization. We can also move some of the blending code to the pattern rather than the mask, which is also clearer I think.

### Submitter Checklist
- [x] Make sure you are opening from a topic branch (**/feature/fix/docs/ branch** (right side)) and not your main branch.
- [x] Ensure that the pull request title represents the desired changelog entry.
- [x] New public fields and methods are annotated with `@since TODO`.
- [x] I read and followed the [contribution guidelines](https://github.com/IntellectualSites/.github/blob/main/CONTRIBUTING.md).