知识层模型(KLM):证据可见的 AI 输出、路由诊断与负结果
The Knowledge Layer Model (KLM): Evidence-Visible AI Output, Routing Diagnostics, and Negative Results
Omer Asim Ozturk · Cihan Aydiner
中文摘要
AI 答案不能只靠表面正确:审阅者需要检查其记录的依据。知识层模型(KLM)是一套开放、与实现无关的推理记录规范,包含带版本的认知评估、证据关系和显式未知项。SEDIM 是 Nage AI OS 的知识与证据架构,作为参考实现。引用 KLM v0.3(仓库标签 v0.3.0)。实证材料为若干独立内部实验,并非端到端验证。一个历史平台域路由基准在 15 个类别上报告 137/150 次分类(91.3%;名义 95% Wilson 区间 85.7-94.9%);分子由四舍五入的汇总反推,历史样例归属未核实。一个原生模型权重路径配置在修正的双源测试中仅 0/15 成功作答。容量增长对比因旧评分器缺陷和原始输出缺失而无法确立可靠效应。词法与作用域诊断对静默与矛盾样例的 AUC 均为 0.008,低于注册下限 0.65。旧打包在 118/221 个选中案例中丢失含金标文本;修复后打包在 221/221 中保留。这些只衡量文本存在性,不衡量答案质量。Python 与 TypeScript 验证器据称在每个合成夹具的六项检查上一致;另一项生产披露未被接受。恢复、重推导、人工评分支持、边界隔离与有界响应评估仍属提议。发现为内部且由作者报告;历史观察与后续复现相区分。证据可见性不确立正确性、因果依赖、安全性或法规合规。独立复现与完整公开审计包仍属未来工作。
关键要点
- 01KLM 是开放、与实现无关的推理记录规范,含版本化认知评估、证据关系与显式未知项。
- 02历史路由基准 137/150(91.3%),但分子由四舍五入汇总反推,样例归属未核实。
- 03原生权重路径配置在双源测试中 0/15 成功;容量增长对比因评分器缺陷无法确立效应。
- 04词法与作用域诊断 AUC 均为 0.008,远低于注册下限 0.65;修复打包保留 221/221 金标文本。
- 05证据可见性不确立正确性、因果依赖、安全性或合规;独立复现与公开审计包仍待完成。
解读
尚无解读。
原始英文摘要
arXiv:2610.42066v1 Announce Type: cross Abstract: AI answers require more than apparent correctness: reviewers need to inspect their recorded basis. We introduce the Knowledge Layer Model (KLM), an open, implementation-neutral specification for inference records with versioned epistemic assessments, evidence relations, and explicit unknowns. SEDIM, the knowledge and evidence architecture of Nage AI OS, is a reference implementation. We reference KLM v0.3 (repository tag v0.3.0). Empirical material consists of separate internal experiments, not end-to-end validation. A historical platform-domain routing benchmark reports 137/150 classifications across 15 classes (91.3%; nominal 95% Wilson interval 85.7-94.9%); the numerator is recovered from a rounded summary, and historical example membership is unverified. A native-model weight-path configuration yielded 0/15 successful answers in a corrected two-source test. A capacity-growth comparison could not establish a reliable effect because of a defective legacy scorer and missing original outputs. Lexical and scoped diagnostics each yielded AUC 0.008 for silent-versus-contradicted examples, below the registered minimum of 0.65. Legacy packing lost gold-containing text in 118/221 selected cases; repaired packing retained it in 221/221. These measure text presence, not answer quality. Python and TypeScript verifiers reportedly agreed on six checks per synthetic fixture; a separate production disclosure was not accepted. Restoration, re-derivation, human-rubric support, boundary isolation, and bounded-response evaluations remain proposed. Findings are internal and author-reported; historical observations are distinguished from later reproductions. Evidence visibility does not establish correctness, causal reliance, security, or regulatory compliance. Independent replication and a complete public audit package remain future work.