Compare and contrast different strategies for multimodal alignment between text embeddings and video frames at Mirage.
