ds_support_golden_v4
support-golden-v4
Curated gold questions and answers across all help-center domains, human-verified by the support quality team.
Items
1,240
Human-verified
97%
Gold coverage
95%
Quality score
94/100
Items by domain
10 of 10 Help Center domains
Evaluation history
10 release and experiment runs · 90 nightly runs on a stratified slice
- v3.4 · Claude Haiku 5.5 comparison
Claude Haiku 5.5 · yesterday
82.3% - v3.4 · Gemini 2.5 Pro comparison
Gemini 2.5 Pro · 2 days ago
83.8% - v3.4 · GPT-5 comparison
GPT-5 · 2 days ago
85.3% - support-assistant v3.4 · release candidate
Claude Sonnet 5.5 · 2 days ago
86.9% - v3.4 · Llama 4 Maverick (self-hosted)
Llama 4 Maverick (self-hosted) · 1 week ago
79.3% - v3.4 · Cohere embed-v4 + rerank-3.5
Claude Sonnet 5.5 · 1 week ago
86.1% - v3.4 · Qwen3-235B comparison
Qwen3-235B · 2 weeks ago
80.2%
24 open or triaged failures trace back to this dataset. Latest: F-2412
Recently verified items
39 most recently verified of 1,240 items. Select an item to see its source and edit it.
My bill says 41,20 €. Is that forty-one euros or four thousand? gt_1007 |
Is SIM swap protection turned on by default? gt_1005 |
Does the Fall 2026 upgrade offer apply if I'm on a family plan? gt_4396 |
Why was I charged twice this month? gt_4403 |
If I upgrade my plan mid-month, how is my next bill calculated? gt_1002 |
Does the Fall 2026 upgrade offer work with a family plan discount? gt_1009 |
What is the excess on a theft insurance claim? gt_1060 |
Give me the steps as a numbered list, please: how do I reset my network settings on Android? gt_4379 |
Why was I charged twice for my September bill? gt_1001 |
How many lines can I add to a family plan? gt_4385 |