[Submitted on 24 Jul 2026]

View PDF HTML (experimental)

Abstract:Air-gapped and on-premises deployments in regulated settings (clinical FHIR services, legal review, sovereign infrastructure) cannot call frontier APIs; they run quantized 4-8B models via this http URL or vLLM behind tool servers. The dominant reliability failure is omission: the silent absence of a decision-critical fact, such as an agent reading 20 of 400 records and reporting "no anomalies." We argue omission is a pipeline phenomenon, not a model phenomenon, and make four contributions. First, a nine-layer taxonomy (L0-L8) locating every omission mechanism from ingestion through the agent loop. Second, an attribution methodology separating deterministic layers (L0-L3) from behavioral layers (L4-L8) via controlled ablation and logit decomposition, quantifying each with an omission waterfall. Third, an open cross-architecture harness comparing sliding-window-hybrid, full-attention, and SSM-hybrid models across engines and frameworks. Fourth, a runtime-detection framework for air-gapped settings where you own the logits. Results from a 75,476-trial sweep across five models and two engines show a pooled omission rate of 0.62; 68% originates in deterministic middleware (L0-L3), relocating where operators should intervene. Server-side profile factors (weight quantization, KV-cache type, RoPE scaling) were fixed and left for future work.

Submission history

From: Santhiya Rajan [view email]
[v1] Fri, 24 Jul 2026 16:10:21 UTC (37 KB)