Seoul National University's PN-Mix Boosts Low-Bit KV Cache Quantization via Pair-Norm Guidance
snu · hf · 2026-10-16
Researchers at Seoul National University introduce PN-Mix, a mixed-precision key cache quantization method for long-output LLM inference. By identifying salient RoPE channel pairs through query-key norms and attention logit perturbation analysis, it assigns higher precision to critical channels, maintaining strong accuracy at low bit rates where standard low-bit quantization severely degrades performance.
More from Research
- Is math just a language? A debate on why LLMs are good at mathematics — LucaAmb · 2026-10-16
- Alibaba open-sources OpenGCUA, a full-stack generalist computer-use agent; +19.6 points on OSWorld-v2 across 8 models — Tongyi-MAI · 2026-10-16
- CoDA co-evolves simulation and robot policies via shared failure memory, lifting OOD success from 5.3% to 27.6% — PLAN-Lab · 2026-10-16
- Cambridge NLIP seminar: Danae Sanchez Villegas on multimodal reasoning beyond final answers — nikaletras · 2026-10-16
- DKGI paper: agents write their own tools to search knowledge graphs, beating baselines by 6.7 Hit@1 — yuz9yuz · 2026-10-16
- ETH Zurich's LFD lets a frozen LLM teach drug-design generators natural-language goals — rohanpaul_ai · 2026-10-16