Comment by SubiculumCode

Are any of these multimodal yet? I'd love to try asking a model with calibrated probabilities to answer question like, "do these shapes match?". Sure, you can ask a LLM....