open/ms-ai-architect
1
0
Fork 0
Code Issues Releases 1 Activity

fix(ms-ai-architect): Spor 0 — 37 kilde-verifiserte KB-feil fikset (25 ref-filer), 1 avvist [skip-docs]

Browse source 
Per-kilde verifiserings-agenter (Opus xhigh, live microsoft_docs_fetch) bekreftet
hver verdi mot kilden FOER endring; alle forekomster av samme gale fakta fikset
fil-vidt (ikke bare sitert linje).

- 37/38 confirm-fix anvendt; #8 (model-selection «Model Router GA») AVVIST: fila var
  allerede korrekt (Model Router GA siden 2025-11-18); den siterte model-choice-guide
  har utdatert «(preview)»-etikett. AA «fikse» ville innfoert en feil.
- Tverteklynger reconciled til kilde-sann verdi: AI Search storage (S1 160/S2 512/S3
  1024/L1 2048/L2 4096 GB; vektor 5/35/150/300), Quota Tiers (erstatter
  Default/Enterprise + «1 Unit Capacity»).
- Hoey-innsats: realtime Schrems II omskrevet (global deployment != EU-residens,
  selv-verifisert mot kilde); Data Zone Norway East = gpt-5.4 OG gpt-5.5 (ikke kun
  5.5); computer-use = gpt-5.4 + computer-tool (region flagget for verifisering).
- Suite 552/552 groenn. Manifest oppdatert (fixed/verdict/verified per fiks).

Spor 1-froe (cross-fil-gjentakelser i UBEROERTE filer): «DDoS Protection Standard» i
ros-ai-threat-library.md + zero-trust-ai-services.md. Tilstoetende funn (ikke i de 38):
se docs / STATE.
This commit is contained in:
Kjell Tore Guttormsen 2026-06-29 09:39:20 +02:00
commit 2c54f0d5a0
26 changed files with 327 additions and 207 deletions
Download patch file Download diff file Expand all files Collapse all files

2
skills/ms-ai-engineering/references/mlops-genaiops/genaiops-llm-specific-practices.md
View file

@ -143,7 +143,7 @@ MLflow Tracing provides end-to-end observability for GenAI applications:
**Hva:** Unified platform for GenAI lifecycle management.
**GenAIOps capabilities:**
- **Model Catalog**: Browse 1600+ foundation models (OpenAI, Meta, Mistral, Cohere)
- **Model Catalog**: Browse over 1,900 models (OpenAI, Meta, Mistral, Cohere)
- **Prompt Flow**: Visual designer for LLM workflows
- **Evaluation SDK**: Built-in evaluators (groundedness, relevance, coherence, fluency, safety)
- **Content Safety**: Real-time filtering (hate, violence, sexual, self-harm)

5
skills/ms-ai-engineering/references/mlops-genaiops/llm-evaluation-production.md
View file

@ -579,9 +579,6 @@ MLflow 3 (SDK `mlflow[databricks]>=3.1`) introduces a unified evaluation model:
| `RetrievalGroundedness` | No | Hallucination detection |
| `Safety` | No | Harmful/toxic content |
| `Correctness` | Yes | Accuracy vs ground truth |
| `Completeness` | Yes | All questions addressed |
| `Fluency` | No | Grammatically correct and naturally flowing |
| `Equivalence` | Yes | Response equivalent to expected output |
| `RetrievalSufficiency` | Yes | Context provides all necessary information |
| `ToolCallCorrectness` | Yes | Tool calls and arguments |
| `ToolCallEfficiency` | No | Redundant tool usage |
@ -1112,7 +1109,7 @@ Dette området utvikler seg raskt. Anbefalt re-verification:
- **Annually:** Compliance requirements (AI Act implementation guidance evolves)
**Siste research-dato:** 2026-06-19
**Kilder brukt:** 7 Microsoft Learn articles, 15 code samples, Azure AI Evaluation SDK v1.14.0
**Kilder brukt:** 7 Microsoft Learn articles, 15 code samples, Azure AI Evaluation SDK v1.17.0
---