Inference Stack Degradation Theory (v1.5)

An author-published research proposal about load-coupled numerical drift in LLM inference and monitoring through a Prompt Weather Index. It distinguishes stability from correctness and sets out a validation path.

Inference Stack Degradation Theory Load-Coupled Numerical Drift and Monitoring in LLM Inference Version 1.5 Roble Mumin Independent AI Researcher and AI Consultant https://roblemumin.com December 2025 Inference Stack Degradation Theory v1.5 Contents 1 Introduction 2 Scope and Non-Claims 3 Established Findings from Prior Work 3.1 Floating-Point Non-Associativity 3.2 Batch Dependence as Practical Cause 4 Relation to ML Systems Research 5 Inference Stack Degradation Theory 5.1 Formal Framing 5.2 Null Hypothesis 6 Autoregressive Sensitivity Considerations 7 Prompt Weather Index 7.1 Motivation 7.2 Definition 7.3 Stability Versus Correctness 7.4 Falsifiability 8 Minimal Empirical Validation Path 9 Operational and Ethical Considerations 10 Implications if Validated 11 Open Research Questions 12 Conclusion Roble Mumin Page 2 of Inference Stack Degradation Theory v1.5 Experiments 13 Experiments 13.1 Experiment 1: Reproducing Batch-Induced Token-Level Divergence 13.1.1 Objective 13.1.2 Setup 13.1.3 Expected Metrics 13.1.4 Null Hypothesis 13.1.5 Agentic Execution Notes 13.2 Experiment 2: Measuring Autoregressive Drift Accumulation 13.2.1 Objective 13.2.2 Setup 13.2.3 Expected ISDT Outcome 13.2.4 Agentic Execution Notes 13.3 Experiment 3: Prompt Weather Index (PWI) Time-Series Monitoring 13.3.1 Objective 13.3.2 Setup 13.3.3 Null Hypothesis 13.3.4 Elaboration: Metric Sensitivity and Weight Robustness 13.3.5 Agentic Execution Notes 13.4 Experiment 4: Task-Level Correctness…
Open Document
Cite this document

Roble Mumin “Inference Stack Degradation Theory (v1.5)”. 2025-12. Available: Document landing page

current summary qualified; independent peer review not verified

Search and filter the archive

Search by title, description, or category. Every entry is a concrete artifact of the system — a paper, a framework, or an applied study.

Podcast RSS
View:

Archive categories

Each category shows a different face of the same method. Research threads feed papers; papers consolidate into long-form works; applied studies prove the pipeline on real problems.