Jet-QSparse: Accelerating Agentic LLM Inference with 4-Bit Sparse Attentionwith Haozhan Tang, Hao Kang, Yecheng Wu, Zerui Wang +51 versionW_pta77t4gv1 · current