I think you’re mystifying a lot of people in this thread! Can you add to your explanation why it’s hard for ML to associate the negation with “stripes”? It seem

I think you’re mystifying a lot of people in this thread! Can you add to your explanation why it’s hard for ML to associate the negation with “stripes”? It seems easy: language => English, in English “without” modifies following phrase, not preceding.