A community-solar grant screener systematically scores one neighborhood lower. How should a tester define MLS bias here?

MLS bias here means non-random, unfair differences on sensitive attributes, like systematically scoring one neighborhood lower. That is not the neuron’s additive bias term on a hidden node, not a random gap with no sensitive attribute, and not a Chapter 6 overfitting finding. Testers hunt unfair treatment, not a math offset.