Runtime-Certified Bounded-Error Quantized Attention
Dean Calver
- Year
- 2026
- Access
- Open access
Abstract
KV cache quantization reduces the memory cost of long-context LLM inference, but introduces approximation error that is typically validated only empirically. Existing systems rely on average-case robustness, with no mechanism to detect or recover from failures at runtime. We present a tiered KV cache architecture that enables runtime-certified attention: INT8 keys and INT4 values are stored in GPU memory, while FP16 originals are retained in system RAM for deterministic fallback. A two-term error decomposition yields per-head, per-step bounds on (i) attention distribution distortion from key quantization and (ii) value reconstruction error. These bounds are computed online and used to drive adaptive precision selection and a multi-stage fallback ladder, which guarantees recovery to the exact dense attention output when required. Across PG-19, NIAH, and RULER benchmarks on LLaMA~3.1-8B with contexts up to 128K, the system matches dense FP16 KV quality within noise for language modelling and retrieval tasks, while recovering catastrophic failures observed in naive INT8/INT4 baselines. Value-sensitive tasks at short context expose a controlled trade-off between compression and fidelity, which can be eliminated via tighter value tolerances or FP16-value fallback. The certification is local (per-head, per-step) and does not guarantee end-to-end model correctness, but ensures that each attention computation is either bounded relative to an FP16 reference or exactly recovered via fallback. This reframes KV cache quantization as a runtime-verified computation rather than a fixed approximation. The goal is not raw speedups, but enabling safe deployment of aggressive KV compression under strict quality constraints.
Keywords
Related papers
A dual-loop framework for manufacturability-aware topology optimization of electric vehicle structures via wire arc additive manufacturing
Qiang Cui, Chuan Yu, Daoqian Yang +2 more
Robotics and Computer-Integrated Manufacturing · 2026
Geometric digital twin: A digital and intelligent model for aero-engine assembly accuracy prediction
Ke Shang, Xin Jin, Teli Xu +4 more
Robotics and Computer-Integrated Manufacturing · 2026
Revolutionizing Industries Through AI-Driven Robotics
Aryan Chaudhary
Recent Advances in Computer Science and Communications · 2026
Design and dynamic performance prediction of a novel large-aperture offset-feed deployable antenna
Chuang Shi, Tianming Liu, Ning Xue +6 more
Aerospace Science and Technology · 2026