News for hackers. Stuff that matters.
Re:Tinted Frames: Question Framing Blinds Vision-Language Mo... by ritwikgupta on Friday March 20, 2026 @01:27PM (#47457785)
VLMs are selectively blind — they decide how much to look at an image based on question framing (open-ended vs Yes/No vs MCQ), even when the same visual reasoning is required.