Amidst the impact of large language models, the fundamental drive to enhance search engines has shifted notably, transitioning from token-based searches to semantic searches. Embeddings are central to this transformation, which play a crucial role in contextualizing searches and facilitating text generation through few-shot transfer.