7 papers
OffQ: Taming Structured Outliers in LLM Quantization by Offsetting
Haoqi Wang, Lorenz K. Mueller, Jiawei Zhuang +2
Low-bit quantization has been widely adopted to accelerate the inference of large language models (LLMs) by significantly reducing computational cost and memory usage. However, act…
KVarN: Variance-Normalized KV-Cache Quantization Mitigates Error Accumulation in Reasoning Tasks
Lorenz K. Muller, Philippe Bich, Chiara Boretti +3
Test-time scaling is a powerful approach to obtain better reasoning in large language models, but it becomes memory-bottlenecked during long-horizon decoding, as the KV-cache grows…
Fast and Stable Triangular Inversion for Delta-Rule Linear Transformers
Aleksandros Sobczyk, Gioele Gottardo, Christos K. Matzoros +4
Linear attention has emerged as a cornerstone for efficient long-context architectures, as evidenced by its integration into state-of-the-art open-source models including Qwen3.5/3…
SINQ: Sinkhorn-Normalized Quantization for Calibration-Free Low-Precision LLM Weights
Lorenz K. Müller, Philippe Bich, Jiawei Zhuang +3
Post-training quantization has emerged as the most widely used strategy for deploying large language models at low precision. Still, current methods show perplexity degradation at…
TyphoonMLA: A Mixed Naive-Absorb MLA Kernel For Shared Prefix
Ahmet Caner Yüzügüler, Ahmet Çelik, Jiawei Zhuang +1
Multi-Head Latent Attention (MLA) is a recent attention mechanism adopted in state-of-the-art LLMs such as DeepSeek-v3 and Kimi K2. Thanks to its novel formulation, MLA allows two…
PRESERVE: Prefetching Model Weights and KV-Cache in Distributed LLM Serving
Ahmet Caner Yüzügüler, Jiawei Zhuang, Lukas Cavigelli
Large language models (LLMs) are typically served from clusters of GPUs/NPUs that consist of large number of devices. Unfortunately, communication between these devices incurs sign…
SSSD: Simply-Scalable Speculative Decoding
Michele Marzollo, Jiawei Zhuang, Niklas Roemer +3
Speculative Decoding has emerged as a popular technique for accelerating inference in Large Language Models. However, most existing approaches yield only modest improvements in pro…