Showing 2026Show all
2 papers · 1 filter
cs.CL2026
STEPQuant: When and Where Errors Matter in Delta-Rule Recurrent State Quantization
Bingchen Yao, Haobo Xu, Haokun Lin +6
Linear attention replaces growing KV caches with fixed-size recurrent states, yet these persistent states can become a substantial memory bottleneck under concurrent serving. Direc…
cs.CV2026
DuQuant++: Fine-grained Rotation Enhances Microscaling FP4 Quantization
Haokun Lin, Xinle Jia, Haobo Xu +7
The MXFP4 microscaling format, which partitions tensors into blocks of 32 elements sharing an E8M0 scaling factor, has emerged as a promising substrate for efficient LLM inference,…