ENTRAP-VL: A Taxonomic Probe for Dual Contextual Entrainment in Vision-Language Models Paper • 2607.20092 • Published Jul 22 • 2 • 3