Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference

Tang, Jiaming, Zhao, Yilong, Zhu, Kan · arXiv · 2024