Publication

Direction is what you need: Improving Word Embedding Compression in Large Language Models