run_2026-07-14_nightly
Nightly · v3.2 · 2026-07-14
Claude Sonnet 5.5 · bge-m3 · dense · support-prompt v5 on support-golden-v4 v4.6
Stratified slice of support-golden-v4 (411 items).
Pass rate
77.6%
Faithfulness
0.86
Answer correctness
0.82
Context precision
0.70
Context recall
0.74
Latency p95
3.3s
Failures by category
92 failed samples · largest share: unsupported
| Category | Share | vs reference |
|---|---|---|
| Unsupported claims | 23.9% | +0.3 pts |
| Incorrect factual answers | 19.6% | +0.2 pts |
| Missing context | 22.8% | +0.2 pts |
| Retrieval mismatch | 23.9% | +0.3 pts |
| Instruction-following failures | 6.5% | −1.0 pts |
| Other | 3.3% | 0.0 pts |
Run details
- Dataset
- support-golden-v4 v4.6 · 411 items
- Generator
- Claude Sonnet 5.5
- Retriever
- bge-m3 · dense
- Prompt
- support-prompt v5
- Judge
- LLM judge · Claude Opus 5.5 · rubric v3
- Started
- Jul 14, 02:00 UTC
- Duration
- 45m 38s
- Triggered by
- Scheduler
- Tokens / request
- 3,870 in · 230 out
- Compared with
- Nightly · v3.2 · 2026-07-13
- nightly
- golden-slice
All metrics
Against Nightly · v3.2 · 2026-07-13
| Metric | Group | This run | Reference | Change |
|---|---|---|---|---|
| Pass rateShare of samples the judge marked as passing all rubric checks. | Outcome | 77.6% | 77.1% | +0.5 pts |
| FaithfulnessShare of answer claims supported by the retrieved context. | Generation | 0.86 | 0.85 | +0.01 |
| Answer correctnessAgreement with the gold answer on facts, not style. | Generation | 0.82 | 0.81 | +0.01 |
| Instruction adherenceRespect for tone, format, length, language and policy constraints. | Generation | 0.88 | 0.87 | 0.00 |
| Context precisionRank-weighted precision of relevant chunks passed to the generator. | Retrieval | 0.70 | 0.70 | 0.00 |
| Context recallShare of gold-answer statements supported by the retrieved context. | Retrieval | 0.74 | 0.75 | −0.01 |
| Recall@5Share of gold passages found in the top 5 retrieved chunks. | Retrieval | 0.78 | 0.78 | 0.00 |
| MRRMean reciprocal rank of the first relevant chunk. | Retrieval | 0.63 | 0.63 | 0.00 |
| NDCG@10Graded ranking quality of the top 10, normalized by the ideal order. | Retrieval | 0.73 | 0.72 | 0.00 |
| Latency p50Median end-to-end time to final token. | Operational | 1.7s | 1.7s | 0.0s |
| Latency p9595th percentile end-to-end time to final token. | Operational | 3.3s | 3.5s | −0.2s |
| Cost per requestModel, embedding and reranking cost per answered request (USD). | Operational | $0.0045 | $0.0045 | 0.0000 |
Nightly trend
Nightly runs on the support-golden-v4 slice, Jul 11 to Oct 8. This run is one point in the series.
Across datasets
support-assistant v3.2 on every dataset in the project. Rows marked “matrix” come from the comparison matrix rather than a dedicated run.
| Dataset | Items | Pass rate | Faith. | Ctx recall | Recall@5 | p95 | Source |
|---|---|---|---|---|---|---|---|
| support-golden-v4this run | 1,240 | 77.6% | 0.86 | 0.74 | 0.78 | 3.3s | This run |
| billing-edge-cases | 312 | Not evaluated with this configuration | |||||
| roaming-and-travel | 188 | Not evaluated with this configuration | |||||
| device-troubleshooting | 426 | Not evaluated with this configuration | |||||
| account-security | 154 | Not evaluated with this configuration | |||||
| locale-formatting | 96 | Not evaluated with this configuration | |||||
| synthetic-hard-negatives | 540 | Not evaluated with this configuration | |||||
Traced samples
17 samples with full traces (4 failing, 13 passing), stratified by domain. Open a row to inspect retrieval and the judge's reasoning.
Can I return my phone after 20 days if I don't like it? Returns & warranty FailIncorrectjudge 0.11 |
What happens if I go over my fair-use limit on Unlimited at home? Plans & upgrades FailMissingjudge 0.07 |
What is the excess on a theft insurance claim? Returns & warranty Passjudge 0.97 |
Does my insurance cover a stolen phone? Returns & warranty FailUnsupportedjudge 0.09 |
My eSIM QR code says it has already been used. What should I do? eSIM & activation Passjudge 0.91 |
Is SIM swap protection turned on by default? Account & security Passjudge 0.95 |
Can I return a phone I bought online if I opened the box? Returns & warranty Passjudge 0.87 |
My bill says 41,20 €. Is that forty-one euros or four thousand? Billing & payments Passjudge 0.96 |
Why was I charged twice for my September bill? Billing & payments Passjudge 0.98 |
If I upgrade my plan mid-month, how is my next bill calculated? Plans & upgrades Passjudge 0.90 |