Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient Language Models Paper โข 2402.19427 โข Published Feb 29, 2024 โข 56