H2EAL: Hybrid-Bonding Architecture with Hybrid Sparse Attention for Efficient Long-Context LLM Inference
Paper • 2508.16653 • Published • 1
Zizhuo Fu
pkuzzz
AI & ML interests
None yet
Recent Activity
upvoted a paper about 6 hours ago
H2EAL: Hybrid-Bonding Architecture with Hybrid Sparse Attention for Efficient Long-Context LLM Inference upvoted a paper about 6 hours ago
Attention Sink Forges Native MoE in Attention Layers: Sink-Aware Training to Address Head Collapse upvoted a paper about 6 hours ago
Scaling Parameter and Context in Attention: Native Sparse Attention from Mixture-of-HeadOrganizations
None yet
upvoted 4
papers about 6 hours ago
Attention Sink Forges Native MoE in Attention Layers: Sink-Aware Training to Address Head Collapse
Paper • 2602.01203 • Published • 1
Scaling Parameter and Context in Attention: Native Sparse Attention from Mixture-of-Head
Paper • 2609.38832 • Published • 1
MAServeBench: Benchmarking Multi-Agent LLM Serving Systems with Hierarchical Collaboration Workloads
Paper • 2610.40293 • Published • 9