Kjell Tore Guttormsen
79e34312ea
Blind fan-out: 15 batcher Opus 4.8 xhigh, 255 P-påstander dekket (0 missing/dupes/extra), hver judget mot live MS Learn via microsoft_docs_fetch (judgen så aldri gull-verdict). Resultat mot forhåndsregistrert gate (recall >=0.80, presisjon >=0.70): - judge: presisjon 81.3% (PASS), recall 68.4% (FAIL, Wilson [52.5-80.9%]), F1 0.743, 26/38 fanget - staleness baseline: recall 0/38 (judge slår den desisivt) - recall-drag konsentrert: sku 37.5%, taxonomy 66.7%; version/status/tpm allerede 80-86% - 10/12 bommer var grounded-men-feil (brittle dekomponering, isolert til sku/taxonomy) Per pre-registrert gate: STOPP, bygg IKKE S3. Fork (operatør) registrert i plan-doc: (a) honorer FAIL, (b) scoped judge, (c) målrettet iterasjon + re-kjør.
1376 lines
59 KiB
JSON
1376 lines
59 KiB
JSON
{
|
|
"_meta": {
|
|
"source": "blind judge fan-out (15 batches, Opus 4.8 xhigh, microsoft_docs_fetch)",
|
|
"population": "volatile + fetchable claim_type (price excluded)",
|
|
"blind": "judges never saw gold verdict; joined by id in run-judge-bakeoff.mjs",
|
|
"note_evidence": "per-claim evidence_quote omitted to bound size; reasons + full quotes in session transcript (same pattern as gold set)",
|
|
"claim_count": 255
|
|
},
|
|
"results": [
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/machine-learning/prompt-flow/migrate-prompt-flow-to-agent-framework",
|
|
"reason": "retirement date April 20, 2027 matches 2027-04-20"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/databricks/mlflow3/genai/eval-monitor/concepts/judges/",
|
|
"reason": "ten single-turn built-in judges match"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#3",
|
|
"judge_verdict": "source_silent",
|
|
"evidence_url": "https://learn.microsoft.com/azure/databricks/mlflow3/genai/eval-monitor/concepts/judges/",
|
|
"reason": "Completeness/Fluency/Equivalence not listed on any learn page"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/databricks/mlflow3/genai/eval-monitor/concepts/judges/",
|
|
"reason": "seven multi-turn conversational judges match"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/concepts/evaluation-evaluators/risk-safety-evaluators",
|
|
"reason": "violence/sexual/self-harm/hate on 0-7 scale confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/concepts/evaluation-evaluators/general-purpose-evaluators",
|
|
"reason": "Groundedness/Relevance/Coherence on 1-5 Likert confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#7",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/python/api/overview/azure/ai-evaluation-readme",
|
|
"reason": "claim SDK v1.14.0 but current is 1.17.0 (superseded)"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/concepts/evaluation-evaluators/agent-evaluators",
|
|
"reason": "Intent Resolution/Task Adherence/Tool Call Accuracy listed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#9",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/databricks/mlflow3/genai/eval-monitor/concepts/scorers",
|
|
"reason": "Databricks-hosted judges, EU models, abuse-monitoring opt-out confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/llm-evaluation-production.md#10",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/databricks/mlflow3/genai/eval-monitor/production-monitoring",
|
|
"reason": "max 20 scorers per experiment confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/what-is-foundry",
|
|
"reason": "Microsoft Foundry rebrand confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#2",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/what-is-foundry",
|
|
"reason": "~1,900 is total catalog not the sold-by-azure subset; 11,000+ total not stated"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/reasoning",
|
|
"reason": "gpt-5 2025-08-07 400k, gpt-5.5 2026-04-24 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
|
|
"reason": "gpt-4.1 2025-04-14 1,047,576 context confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
|
|
"reason": "DeepSeek-R1 163,840 context confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#6",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
|
|
"reason": "claim 131,072 but DeepSeek-R1-0528 is 163,840"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#7",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry-classic/openai/how-to/computer-use",
|
|
"reason": "page now gpt-5.4 with new access link; computer-use-preview/2025-03-11/8192/1024 gone"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/azure-ai-foundry.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/what-is-foundry",
|
|
"reason": "SDK Python/C#/JS/Java confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
|
|
"reason": "gpt-5 2025-08-07 400k/272k/128k matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
|
|
"reason": "gpt-5.5 2026-04-24 1,050,000/922k/Dec2025 matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/provisioned-throughput-sizing",
|
|
"reason": "gpt-5 4,750 / gpt-5-mini 23,750 Input TPM per PTU matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#4",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/provisioned-throughput-sizing",
|
|
"reason": "claim ratio 1 but page states DeepSeek output-to-input ratio 4"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure-region-availability",
|
|
"reason": "only gpt-4o/gpt-4o-mini in norwayeast residency-bound tables"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#6",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure-region-availability",
|
|
"reason": "gpt-5.4 also Data Zone Standard norwayeast; not only gpt-5.5"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#7",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/reasoning",
|
|
"reason": "page says access no longer restricted; registration claim contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure-region-availability",
|
|
"reason": "o3-deep-research Global Standard westus+norwayeast only"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/platforms/model-catalog-2026.md#9",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/provisioned-throughput-sizing",
|
|
"reason": "global/data-zone min 15 incr 5, regional 25-50 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/reliability/reliability-api-management",
|
|
"reason": "Premium v2 AZ not multi-region; multi-region only Premium classic"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/deployment-types",
|
|
"reason": "Global/DataZone/Standard residency taxonomy matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/api-management-howto-deploy-multi-region",
|
|
"reason": "<10s config sync, mgmt plane+dev portal primary only"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/api-management-howto-deploy-multi-region",
|
|
"reason": "regional DNS pattern matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/genai-gateway-capabilities",
|
|
"reason": "semantic caching via Managed Redis + llm-semantic-cache policies"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/genai-gateway-capabilities",
|
|
"reason": "llm-emit-token-metric with custom dimensions confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#7",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/genai-gateway-capabilities",
|
|
"reason": "regional gateways + circuit breaker + priority load balancing"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/api-management-howto-deploy-multi-region",
|
|
"reason": "internal VNet no auto-route, port 3443 mgmt, 1433 outbound SQL"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/api-management/multi-region-ai-gateway-design.md#9",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/api-management-howto-deploy-multi-region",
|
|
"reason": "rate-limit/llm-token-limit count separately per regional gateway"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/how-to/high-availability-resiliency",
|
|
"reason": "Foundry no automatic failover/DR confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/how-to/high-availability-resiliency",
|
|
"reason": "conversation history in Cosmos DB enterprise_memory confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/how-to/high-availability-resiliency",
|
|
"reason": "Cosmos continuous backup 30-day PITR tier confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/how-to/high-availability-resiliency",
|
|
"reason": "GZRS for Storage account prescribed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/how-to/high-availability-resiliency",
|
|
"reason": "Cosmos Service-Managed Failover switches write region"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/how-to/high-availability-resiliency",
|
|
"reason": "Key Vault auto-failover confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#7",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/how-to/high-availability-resiliency",
|
|
"reason": "ACR geo-replication instruction confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/what-is-foundry",
|
|
"reason": "Azure AI Studio -> Microsoft Foundry brand mapping confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/ai-foundry-disaster-recovery-planning.md#9",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/ai-foundry/openai/concepts/models",
|
|
"reason": "Global training Public Preview, no residency, Norway East+Sweden Central"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/gpt-with-vision",
|
|
"reason": "10 images per chat request confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/gpt-with-vision",
|
|
"reason": "low=512x512, high=512x512 segments confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/ai-foundry/openai/overview",
|
|
"reason": "GPT-4o low flat rate 85 tokens confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#4",
|
|
"judge_verdict": "source_silent",
|
|
"evidence_url": "",
|
|
"reason": "high-detail totals 4160/6240/6208 on no learn page (evidence_url null)"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/gpt-with-vision",
|
|
"reason": "GPT-5/4.1/4.5/4o + o-series vision-enabled match"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/multimodal-search-overview",
|
|
"reason": "Document Extraction + Content Understanding recommended; Document Layout legacy"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#7",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/multimodal-search-overview",
|
|
"reason": "AML skill also valid; exclusivity to Azure Vision contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/multimodal-prompt-design.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/multimodal-search-overview",
|
|
"reason": "Content Understanding cross-page tables/chunking/PDF-DOCX-XLSX-PPTX confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/responsible-ai/algorithmic-accountability-auditability.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/purview/ai-microsoft-purview",
|
|
"reason": "Enterprise AI apps taxonomy (Foundry/Claude Enterprise/ChatGPT Enterprise) matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/cost-models.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/microsoft-copilot-studio/requirements-messages-management",
|
|
"reason": "agent flow enforcement blocks new runs confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/cost-models.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/ai-builder/credit-management",
|
|
"reason": "unused capacity doesn't carry over confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-region-support",
|
|
"reason": "Norway East semantic ranker blank confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-region-support",
|
|
"reason": "Norway East agentic retrieval blank confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-region-support",
|
|
"reason": "Norway East query rewrite blank confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-region-support",
|
|
"reason": "Norway East confidential computing yes, Sweden Central no"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-region-support",
|
|
"reason": "Sweden Central agentic/semantic/query-rewrite yes"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-region-support",
|
|
"reason": "all three regions availability zones yes (3 AZ)"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#7",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/reference/region-support",
|
|
"reason": "Norway East in Foundry projects region list"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/regional-availability-verification.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/ai-foundry/openai/concepts/models",
|
|
"reason": "o3/gpt-4.1/gpt-4o norwayeast yes; tts swedencentral only"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-sku-tier",
|
|
"reason": "Serverless preview 3 regions, no SLA, no tier migration"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-sku-tier",
|
|
"reason": "SU and CU not interchangeable confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-sku-tier",
|
|
"reason": "existing services switch Basic<->S1/S2/S3 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-limits-quotas-capacity",
|
|
"reason": "Basic 15GB from Apr2024, 2GB older confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#5",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-limits-quotas-capacity",
|
|
"reason": "claim S1 25/S2 100/... matches only superseded pre-Apr2024 row; current 160/512/1024/2048/4096"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/search-limits-quotas-capacity",
|
|
"reason": "L1/L2 up to 12 partitions/replicas confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#7",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/vector-search-how-to-quantization",
|
|
"reason": "scalar int8 fourfold (75%) reduction confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-cost-optimization.md#8",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/search/vector-search-how-to-quantization",
|
|
"reason": "page says up to 28x/96%; claim 96.875% (32x) contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/deployment-types",
|
|
"reason": "deployment-type taxonomy confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/backends",
|
|
"reason": "up to 30 backends in pool confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/genai-gateway-capabilities",
|
|
"reason": "Anthropic v2/Vertex AI/Bedrock endpoints confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/genai-gateway-capabilities",
|
|
"reason": "unified model API preview confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/genai-gateway-capabilities",
|
|
"reason": "AI gateway in Foundry preview confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/genai-gateway-capabilities",
|
|
"reason": "semantic caching via Managed Redis confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#7",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/api-management/backends",
|
|
"reason": "429 + Retry-After up to 1 day circuit breaker confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-enterprise-architecture.md#8",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/ai-foundry/openai/concepts/models",
|
|
"reason": "Azure OpenAI models available in norwayeast; 'no Norway region' contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/genaiops-llm-specific-practices.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/machine-learning/prompt-flow/migrate-prompt-flow-to-agent-framework",
|
|
"reason": "retirement 2027-04-20 + Agent Framework migration confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/genaiops-llm-specific-practices.md#2",
|
|
"judge_verdict": "not_grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/concepts/foundry-models-overview",
|
|
"reason": "claim 1600+ but page states over 1,900 (superseded)"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#1",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/azure-monitor/reference/supported-metrics/microsoft-cognitiveservices-accounts-metrics",
|
|
"reason": "TokenTransaction = input+output confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#2",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/azure-monitor/reference/supported-metrics/microsoft-cognitiveservices-accounts-metrics",
|
|
"reason": "ProcessedPromptTokens = input confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#3",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/azure-monitor/reference/supported-metrics/microsoft-cognitiveservices-accounts-metrics",
|
|
"reason": "prompt=input completion=output taxonomy confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#4",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry-classic/openai/concepts/use-your-data",
|
|
"reason": "On Your Data two model calls confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#5",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry-classic/openai/concepts/use-your-data",
|
|
"reason": "gpt-35-turbo-16k token estimation table matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#6",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry-classic/openai/concepts/use-your-data",
|
|
"reason": "default 5 docs, 1024 chunk confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#7",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/well-architected/cost-optimization/collect-review-cost-data",
|
|
"reason": "13mo retention, FOCUS pipeline, 90/100/110% thresholds confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/token-usage-tracking-attribution.md#8",
|
|
"judge_verdict": "grounded",
|
|
"evidence_url": "https://learn.microsoft.com/azure/foundry/openai/how-to/provisioned-throughput-sizing",
|
|
"reason": "Llama-3.3-70B 8,450 Input TPM per PTU confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/reasoning-models-o1-o3-optimization.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "o-series 200k/100k context confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/reasoning-models-o1-o3-optimization.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "unsupported params list matches"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/reasoning-models-o1-o3-optimization.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "xhigh only gpt-5.1-codex-max confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/reasoning-models-o1-o3-optimization.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "verbosity/preamble/allowed-tools/lark_tool confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/reasoning-models-o1-o3-optimization.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-5 series no concise summary confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/reasoning-models-o1-o3-optimization.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "o3-mini/o1 markdown re-enable confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/reasoning-models-o1-o3-optimization.md#7",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "o3-pro/gpt-5-pro/gpt-5-codex access no longer restricted; registration claim superseded"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/adr-template.md#1",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "Graph Connectors manage ACL permissions; zero-permission claim contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/adr-template.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Standard no SLA vs Provisioned latency target+dedicated confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/architecture/adr-template.md#3",
|
|
"judge_verdict": "source_silent",
|
|
"reason": "no NOK/PTU price or 1PTU~100k tokens/hr on any learn page"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "regional min/incr 50 (25 mini/nano) confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "global/data-zone min 15 incr 5 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "DeepSeek/Fireworks 100+ PTU minimum confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "new commitment enrollment stopped Aug 1 2024 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "PTU reservation cross-model (OpenAI+DeepSeek) confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#6",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "auto-renew default OFF, not ON; claim contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "disconnected 1yr/web monthly/non-refundable/single-service confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/reserved-capacity-planning.md#8",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-5.1 4750 TPM/PTU ratio 8 incr 50 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/token-optimization-and-efficiency.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "1024 min + 128 increment cache confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/token-optimization-and-efficiency.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "24h extended retention + in_memory default confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/token-optimization-and-efficiency.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "5-10min clear, within 1h confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/token-optimization-and-efficiency.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "prompt_cache_key ~15 rpm threshold confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/real-time-reasoning-performance.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "realtime model versions match"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/real-time-reasoning-performance.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-realtime GA status confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/real-time-reasoning-performance.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "GA /openai/v1 endpoint confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/real-time-reasoning-performance.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "none/semantic_vad/server_vad confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/real-time-reasoning-performance.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "East US 2 + Sweden Central global confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/real-time-reasoning-performance.md#6",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "Realtime API now GA; 'still public preview no SLA' contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/real-time-reasoning-performance.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "GPT-4o mini PTU sizing pairs match"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-context-windows.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-4o 128k context confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-context-windows.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-4.1 1,047,576 context (200k+) confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-context-windows.md#3",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "o3-mini is 200k not 128k; contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-context-windows.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "o1 200k context confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-context-windows.md#5",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "quota now Tier-based; Default/Enterprise superseded"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-context-windows.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "data zone batch 500M/30M quota confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/rag-context-windows.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "File Search max_prompt_tokens 20k/50k confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/monitoring-observability-ml-systems.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "five built-in monitoring signals confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/monitoring-observability-ml-systems.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Jensen-Shannon/chi-squared/NDCG metrics confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/monitoring-observability-ml-systems.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Spark instance types E4s_v3-E64s_v3 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/monitoring-observability-ml-systems.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "event type + filter key confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/monitoring-observability-ml-systems.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "ML namespace + tables confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/monitoring-observability-ml-systems.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "three alert rules aggregation/threshold match"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/mlops-genaiops/monitoring-observability-ml-systems.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "AppTraces 90-day retention confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/multi-region-azure-openai-deployment.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-4o norwayeast confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/multi-region-azure-openai-deployment.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Sweden Central gpt-4o/gpt-35-turbo/o1 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/multi-region-azure-openai-deployment.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "older chat 1U/6RPM/1000TPM confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/multi-region-azure-openai-deployment.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "o1 1U/1RPM/6000TPM confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/multi-region-azure-openai-deployment.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "o3-mini/o1-mini/o3-pro/o3/o4-mini TPM confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/multi-region-azure-openai-deployment.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "30 resources/region, per-model deploy limit removed confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/multi-region-azure-openai-deployment.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "load balancing + session affinity + circuit breaker confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/capacity-planning-dr-configurations.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "ProvisionedManaged SKU name confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/capacity-planning-dr-configurations.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-4o 2024-08-06 version confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/capacity-planning-dr-configurations.md#3",
|
|
"judge_verdict": "source_silent",
|
|
"reason": "99.9%/99.99% per-replica SLA percentages not stated on any learn page"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/capacity-planning-dr-configurations.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "quota region-specific, pre-allocate secondary confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/network-resilience-patterns-ai.md#1",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "tiers are Network/IP Protection; 'Standard' tier name superseded"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/network-resilience-patterns-ai.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "CognitiveServicesManagement service tag confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/network-resilience-patterns-ai.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "AzureCognitiveSearch service tag confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/network-resilience-patterns-ai.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "exponential backoff + circuit breakers confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/service-level-documentation-dr.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "2024-10-21 latest GA api-version confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/service-level-documentation-dr.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "2024-07-01 stable Search API confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/service-level-documentation-dr.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "P2V3/P3V3 App Service SKUs confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/service-level-documentation-dr.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "AI Search single-region, customer dual-index confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/service-level-documentation-dr.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Key Vault MS-managed auto failover confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/service-level-documentation-dr.md#6",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "multi-region writes can't be RPO/RTO=0; <15min RPO; contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/service-level-documentation-dr.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "App Configuration geo-replication recommended confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "model router 2025-11-18 active latest confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#2",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "router supports ~28 models, not 'up to 18'"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#3",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "router rate limits now tier-based; Default/Enterprise values superseded"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-4.1-nano 59,400 TPM/PTU confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-5/gpt-5-mini/o4-mini TPM/PTU confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "East US 2 + Sweden Central router regions confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#8",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Balanced 1-2% / Cost 5-6% quality band confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/gpt5-gpt41-pricing-models.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-5 4750 / gpt-4.1 3000 TPM/PTU confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/gpt5-gpt41-pricing-models.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "output-input ratio gpt-5=8 gpt-4.1=4 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/gpt5-gpt41-pricing-models.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Copilot Studio Basic/Standard/Premium rate tiers confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/gpt5-gpt41-pricing-models.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "four thinking levels + minimal no parallel tools confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/azure-ai-foundry-cost-governance.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Foundry RBAC rename confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/azure-ai-foundry-cost-governance.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Foundry Tools kind=AIServices confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/azure-ai-foundry-cost-governance.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Owner/Contributor vs Foundry User build split confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/azure-ai-foundry-cost-governance.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "no hard limit, budgets alert-only confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/azure-ai-foundry-cost-governance.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "dynamic quota no ceiling confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/azure-ai-foundry-cost-governance.md#6",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "quota now Tier-based; 150K-30M range superseded"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/azure-ai-foundry-cost-governance.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "serverless 200k TPM / 1000 RPM per deployment confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/vector-storage-cost-optimization.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "scalar int8 fourfold confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/vector-storage-cost-optimization.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "binary up to 28x confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/vector-storage-cost-optimization.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "min 1024 dims with binary confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/vector-storage-cost-optimization.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "HNSW 1-20% overhead confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/vector-storage-cost-optimization.md#5",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "post-Apr2024 vector quota 5/35/150/300; claim 1/12/36/72 contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/vector-storage-cost-optimization.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Serverless Developer 3 regions/no SLA/no migration confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/vector-storage-cost-optimization.md#7",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "vector compression GA at 2024-07-01 not 2024-11-01; contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-cost-optimization.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "commitment-tier features list match"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-cost-optimization.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "commitment tier single-service confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-cost-optimization.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Regional/DataZone/Global provisioned types confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-cost-optimization.md#5",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "terms are 1-month/1-year not 1/3-year; contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-cost-optimization.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Anthropic v2/Vertex/unified API/Foundry preview confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-cost-optimization.md#8",
|
|
"judge_verdict": "source_silent",
|
|
"reason": "no learn page states OpenAI lacks hard-limit spend cap (negative)"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-vs-foundry-tools-selection.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Standard/Global Standard/Provisioned types confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-vs-foundry-tools-selection.md#3",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "DALL-E 3 replaced by gpt-image; o4-mini not top reasoning; superseded"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-vs-foundry-tools-selection.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Content Understanding v1.0 GA 2025-11-01 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-vs-foundry-tools-selection.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "over 1,900 models (>100+) confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/azure-ai-services/ai-services-vs-foundry-tools-selection.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Cognitive Services -> Foundry Tools rename confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/small-language-models-economics.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "SLM <10B / LLM >10B threshold confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/small-language-models-economics.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Phi-3-small 7B/medium 14B/Phi-2 2.7B confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/small-language-models-economics.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Phi-4-mini 131072/4096 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/small-language-models-economics.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Phi MIT, Falcon-7B Apache 2.0 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/small-language-models-economics.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Phi-4 sidecar frameworks confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/small-language-models-economics.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "KAITO Phi-4-mini GPU node pool confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/small-language-models-economics.md#9",
|
|
"judge_verdict": "grounded",
|
|
"reason": "A100/T4 region availability confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/embedding-models-selection.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "ada-002 1536/3-small 1536/3-large 3072 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/embedding-models-selection.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "MRL truncation 3-large 256/3-small 512 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/embedding-models-selection.md#4",
|
|
"judge_verdict": "source_silent",
|
|
"reason": "multilingual-e5 dims not on any learn page"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/embedding-models-selection.md#5",
|
|
"judge_verdict": "source_silent",
|
|
"reason": "no 'Custom Models preview' embedding fine-tuning path documented"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/embedding-models-selection.md#6",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "Basic partition 2GB or 15GB, never 1GB; contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-engineering/rag-architecture/embedding-models-selection.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "S1 25GB (pre-Apr2024) confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/endpoint-health-and-capacity-planning.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "AzureOpenAIAvailabilityRate PT1M/no-export confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/endpoint-health-and-capacity-planning.md#2",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "metric names PTUUtilization/ProcessingTime/TokensGenerated wrong; actual ProvisionedUtilization/GeneratedTokens"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/endpoint-health-and-capacity-planning.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "per-unit RPM/TPM ratios match"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/endpoint-health-and-capacity-planning.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "dynamic quota opportunistic, never below base confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/endpoint-health-and-capacity-planning.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Basic 3x3=9 SU confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/endpoint-health-and-capacity-planning.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Serverless CU/hr no SLA 3 regions confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/response-quality-metrics-rag.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "six RAG evaluators listed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/response-quality-metrics-rag.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Fidelity/NDCG/XDCG/MaxRelevance/Holes confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/response-quality-metrics-rag.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "1-5 scale, default threshold 3 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/response-quality-metrics-rag.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Groundedness Pro Content Safety binary confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/response-quality-metrics-rag.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Prompt flow retirement 2027-04-20 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/response-quality-metrics-rag.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "is_reasoning_model o1/o3 config confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/cost-monitoring-cost-attribution.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "project-level chargeback + project tag confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/cost-monitoring-cost-attribution.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "partner models model-name-GUID under Global confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/cost-monitoring-cost-attribution.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "fine-tuned training/hosting/inference billing confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/cost-monitoring-cost-attribution.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "15-day inactive deployment auto-delete confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/monitoring-observability/cost-monitoring-cost-attribution.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "paygo-inference-input-tokens meter names confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-infrastructure/bcdr/cost-analysis-dr-configurations.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "savings plan 1/3-year all-region confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ptu-vs-paygo-economics.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "three provisioned types confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ptu-vs-paygo-economics.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Standard/Priority/Provisioned/Batch categories confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ptu-vs-paygo-economics.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-4o 50/15, gpt-4o-mini 25/15, DeepSeek 100 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ptu-vs-paygo-economics.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-5 4750 TPM/PTU confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ptu-vs-paygo-economics.md#5",
|
|
"judge_verdict": "grounded",
|
|
"reason": "spillover OpenAI-only confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ptu-vs-paygo-economics.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Provisioned-managed Utilization V2 + 429 at 100% confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ai-builder-credits-transition.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "EoS 2025-11-01, EoL/seeded-removal 2026-11-01 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ai-builder-credits-transition.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "seeded credits removed incl EA confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ai-builder-credits-transition.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "per-license credits 500/250/5000/20000/25000 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ai-builder-credits-transition.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "Copilot Studio AI Builder uses Copilot Credits confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ai-builder-credits-transition.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "new customers can't buy add-ons from 2025-11-01 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/ai-builder-credits-transition.md#8",
|
|
"judge_verdict": "grounded",
|
|
"reason": "overage grace period, monthly reset, no carry-over confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/chain-of-thought-prompting.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "reasoning_effort none/minimal/low/medium/high/xhigh confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/chain-of-thought-prompting.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "low/medium/high all except o1-mini confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-advisor/prompt-engineering/chain-of-thought-prompting.md#3",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "gpt-5 series does NOT support concise; claim contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/responsible-ai/stakeholder-communication-ai-decisions.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "RAI scorecard public preview PDF confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/responsible-ai/stakeholder-communication-ai-decisions.md#3",
|
|
"judge_verdict": "grounded",
|
|
"reason": "25,000 Copilot Credits per pack/month confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "o4-mini 5400 / gpt-4.1 3000 TPM/PTU confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#6",
|
|
"judge_verdict": "grounded",
|
|
"reason": "GPT-5 four thinking levels confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "model router up to 60% cost saving confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#8",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "Model Router is preview not GA; contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#9",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "default model now GPT-4.1 mini not GPT-4o-mini; superseded"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/token-counting-optimization.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "1024 min/128 granularity/5-10min TTL confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/token-counting-optimization.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "24h extended cache gpt-5/gpt-4.1 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/token-counting-optimization.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "On Your Data retirement 2026-10-14 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/token-counting-optimization.md#8",
|
|
"judge_verdict": "grounded",
|
|
"reason": "gpt-35-turbo-16k token table + 20% reserve confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/rag-query-cost-reduction.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "On Your Data per-model token sums match"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/rag-query-cost-reduction.md#2",
|
|
"judge_verdict": "grounded",
|
|
"reason": "partition storage Basic2/S1 25/S2 100/S3 200 confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/rag-query-cost-reduction.md#4",
|
|
"judge_verdict": "grounded",
|
|
"reason": "agentic retrieval GA 2026-04-01 REST + Foundry IQ confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/rag-query-cost-reduction.md#7",
|
|
"judge_verdict": "grounded",
|
|
"reason": "ada-002 sole On Your Data embedding confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/responsible-ai/responsible-ai-training-awareness.md#1",
|
|
"judge_verdict": "grounded",
|
|
"reason": "module title/9 Units/Intermediate confirmed"
|
|
},
|
|
{
|
|
"id": "ms-ai-governance/responsible-ai/responsible-ai-training-awareness.md#2",
|
|
"judge_verdict": "not_grounded",
|
|
"reason": "AI-102 has 12-mo renewal + retire 2026-06-30; 'no expiration' contradicted"
|
|
},
|
|
{
|
|
"id": "ms-ai-security/cost-optimization/deterministic-cost-calculation-model.md#9",
|
|
"judge_verdict": "grounded",
|
|
"reason": "messages->Copilot Credits 2025-09-01 switch confirmed"
|
|
}
|
|
]
|
|
}
|