Same Request, Different Answer: Quantization Amplifies Cache-Induced Divergence in LLM Serving

arXiv:2609.04748 · cs.SE, cs.DC, cs.LG · Submitted 2026-09-04 · Read on arXiv

cs.SE, cs.DC, cs.LG

Submitted: 2026-09-04

Updated: 2026-09-04

Comments: Submitted to IEEE Access

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Quantization — 2× in this paper · explained in 12 episodes
• Accuracy — 1× in this paper · explained in 1 episode

Related papers

• Falsification-Based Verification of LLM-Generated Optimization Models: Sound Test Batteries and Their Detection Limits
• GitSkills: A Dataset of Agent Skills on GitHub
• SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
• PackMonitor: Enabling Zero Package Hallucinations Through Decoding-Time Monitoring
• IntentCoding: Amplifying User Intent in Code Generation
• Incentives and Outcomes in Bug Bounties