How Efficient Attention Mechanisms Are Solving the Scalability Crisis in Large Language Models • Libertify