Ontology-Amplified Distillation and Contextuality Auditing for Sovereign Enterprise Language Models: A Combined Proof-of-Mechanism and Negative-Results Method Study

主权企业语言模型的本体增强蒸馏与语境审计：机制验证与负面结果研究

Abstract: Regulated financial institutions operating under data-residency rules need tenant-owned language models that can run inside the institution’s perimeter. This paper combines two related FAOS studies into one mechanism-and-control article. 摘要： 在数据驻留规则下运营的受监管金融机构，需要能够在其内部边界内运行的、由租户自主拥有的语言模型。本文将两项相关的 FAOS 研究合并为一篇机制与控制研究论文。

First, it reports a reduced-power proof-of-mechanism study of ontology-amplified distillation: a Qwen3.6-27B student is adapted to the Foundation AgenticOS ontology through supervised fine-tuning on frontier-teacher trajectories and ontology-grounded direct preference optimization (DPO), trained locally on a single Apple M5 Max from 47 synthetic, English-language, cross-domain preference pairs. 首先，报告了一项关于本体增强蒸馏的低功耗机制验证研究：通过在“前沿教师模型”（frontier-teacher）轨迹上进行监督微调，并结合基于本体的直接偏好优化（DPO），将 Qwen3.6-27B 学生模型适配至 Foundation AgenticOS 本体。该模型在单台 Apple M5 Max 上，利用 47 对合成的英语跨领域偏好数据进行了本地训练。

On 40 held-out Vietnamese financial-domain tasks, the distilled student grounds 36 of 40 tasks (grounded rate 0.90; mean ontology term-coverage r_onto = 0.95 on a metric floored at 0.50), equal to the GPT-5 frontier baseline, which also grounds 36 of 40. The outcome is underpowered to establish equivalence: the paired-difference 95% confidence interval spans +/-4 tasks, and the run does not test or show the pre-registered amplification prediction that the student should exceed the frontier. 在 40 项未参与训练的越南金融领域任务中，蒸馏后的学生模型在 36 项任务中实现了基础支撑（基础支撑率为 0.90；平均本体术语覆盖率 r_onto = 0.95，指标下限为 0.50），与同样完成 36 项任务的 GPT-5 前沿基准模型持平。该结果的统计效力不足以确立等效性：配对差异的 95% 置信区间跨度为 +/-4 项任务，且本次运行未能验证或展示预注册的“增强预测”（即学生模型应超越前沿模型）结论。

Second, the paper consolidates a contextuality-audit method for enterprise-agent routing. In a separate negative-results pilot, the corrected canonical Contextuality-by-Default degree is zero for all Phase 1.3 groups in both the local-Qwen run and an explicitly labeled Gemma replication check; the useful signal is direct influence and construct coupling, not surviving residual contextuality. 其次，本文整合了一种用于企业智能体路由的语境审计方法。在一项独立的负面结果试点研究中，无论是本地 Qwen 运行还是明确标记的 Gemma 复现检查，所有 1.3 阶段组的修正后典型“默认语境性”（Contextuality-by-Default）程度均为零；有效的信号在于直接影响和结构耦合，而非残留的语境性。

Together, the studies pair an ontology-grounded model-building mechanism with a governance diagnostic for deciding when apparent disagreement should trigger prompt standardization, multi-agent synthesis, or human review. The evidence supports neither deployability, safety, superiority, statistical equivalence, nor a contextuality-positive routing rule. 综上所述，这些研究将基于本体的模型构建机制与治理诊断相结合，用于决定何时应针对明显的意见分歧触发提示词标准化、多智能体综合或人工审核。目前的证据既不支持该模型的可部署性、安全性、优越性或统计等效性，也不支持语境正向的路由规则。