Relative Drawing Identification Complexity Is Invariant to Modality in Vision-Language Models | Publicación