RAG — Internal Docs QA

Run run-rag-qa-003 · May 20, 11:15 AM

rubric rag-qa-v2.0model claude-sonnet-4-6dataset rag-golden-set-v2changed: retrieval-topk-4

Verdict

Acceptable

0.79/1.0

Cases

24 passing

Pass rate

80%

Safety findings

Claims processed

Per-case detail not available

This seeded run stores summary totals only — no per-case rows in the public demo dataset. For a fully populated walkthrough (dimension breakdown, claim heat map, evidence chain), open the RAG — Internal Docs QA run.