arxiv:2608.12149
zunhaisu
zunhai
AI & ML interests
None yet
Recent Activity
authored a paper about 11 hours ago
Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus authored a paper about 11 hours ago
SnapMLA: Efficient Long-Context MLA Decoding via Hardware-Aware FP8 Quantized Pipelining authored a paper about 11 hours ago
XStreamVGGT: Extremely Memory-Efficient Streaming Vision Geometry Grounded Transformer with KV Cache CompressionOrganizations
models 0
None public yet
datasets 0
None public yet