VLTest: Finding Joint Image-Text Failures in Vision-Language Modelswith Jintao Lin, Guowei Yang1 versionW_w263sutcv1 · current