Evaluating the Capabilities of Multi-modal Reasoning Models with Synthetic Task Data

Nathan Vaska, Victoria Helus

The impressive advances and applications of large language and joint language-and-visual understanding models has led to an increased need for methods of probing their potential re…