Risk-Controlled KV-Cache Eviction: From Memory Budgets to Risk Targets

arXiv:2609.27981 · cs.CL, cs.LG · Submitted 2026-09-23 · Read on arXiv

cs.CL, cs.LG

Submitted: 2026-09-23

Updated: 2026-09-23

Terminology

• Calibration — 28× in this paper · explained in 7 episodes
• Benchmark — 15× in this paper · explained in 2 episodes
• KV cache — 11× in this paper · explained in 3 episodes
• Adam — 3× in this paper · explained in 2 episodes
• Quantization — 2× in this paper · explained in 12 episodes
• Robustness — 2× in this paper · explained in 5 episodes
• Adaptivity — 1× in this paper · explained in 1 episode
• Aggregation — 1× in this paper · explained in 1 episode
• Attention Entropy — 1× in this paper · explained in 2 episodes
• Distribution Shift — 1× in this paper · explained in 6 episodes
• Few-Shot Learning — 1× in this paper · explained in 4 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Perplexity — 1× in this paper · explained in 3 episodes

Sources

• The Llama 3 Herd of Models
• Mistral 7B
• WitCert: Sound Runtime Risk Observability and Gating for KV-Cache Quantization
• Error Certificates for KV-Cache Eviction via Randomized Design
• CompilerKV: Risk-Adaptive KV Compression via Offline Experience Compilation

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering