In-Context Learning Improves Compositional Understanding of Vision-Language Models was just published in ICML Workshop on Foundation Models in the Wild 🙌