Skip to content

run_2026-07-12_nightly

Nightly · v3.2 · 2026-07-12

Claude Sonnet 5.5 · bge-m3 · dense · support-prompt v5 on support-golden-v4 v4.6

CompletedNightlysupport-assistant v3.2v2 · 512 tokens / 64 overlap
Compare

Stratified slice of support-golden-v4 (423 items).

Pass rate

76.6%

−1.9 pts(regression)vs previous nightly324 of 423 samples passed

Faithfulness

0.85

0.00(no change)vs previous nightly

Answer correctness

0.81

+0.01(improvement)vs previous nightly

Context precision

0.70

+0.01(improvement)vs previous nightly

Context recall

0.75

−0.01(regression)vs previous nightly

Latency p95

3.6s

+0.1s(regression)vs previous nightly

Failures by category

99 failed samples · largest share: unsupported

Share of failed samples by category, compared with Nightly · v3.2 · 2026-07-11
CategorySharevs reference
Unsupported claims24.2%−0.2 pts
Incorrect factual answers19.2%+0.3 pts
Missing context23.2%−0.1 pts
Retrieval mismatch23.2%−0.1 pts
Instruction-following failures7.1%+0.4 pts
Other3.0%−0.3 pts

Run details

Dataset
support-golden-v4 v4.6 · 423 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 12, 02:00 UTC
Duration
40m 59s
Triggered by
Scheduler
Tokens / request
3,870 in · 230 out
  • nightly
  • golden-slice

All metrics

Against Nightly · v3.2 · 2026-07-11

All metrics for Nightly · v3.2 · 2026-07-12
MetricThis runReferenceChange
Pass rate76.6%78.5%−1.9 pts
Faithfulness0.850.850.00
Answer correctness0.810.80+0.01
Instruction adherence0.880.86+0.01
Context precision0.700.69+0.01
Context recall0.750.76−0.01
Recall@50.770.78−0.01
MRR0.630.640.00
NDCG@100.720.73−0.01
Latency p501.7s1.8s0.0s
Latency p953.6s3.4s+0.1s
Cost per request$0.0043$0.0044−0.0001

Nightly trend

Nightly runs on the support-golden-v4 slice, Jul 11 to Oct 8. This run is one point in the series.

Pass rate moved from 78.5% to 87.2% across 90 nightly runs. Configuration changes: v3.3 on Aug 4, v3.4 on Sep 8.

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.

Configuration results per dataset
DatasetItemsPass rateFaith.Ctx recallRecall@5p95Source
support-golden-v4this run1,240
76.6%
0.850.750.773.6sThis run
billing-edge-cases312Not evaluated with this configuration
roaming-and-travel188Not evaluated with this configuration
device-troubleshooting426Not evaluated with this configuration
account-security154Not evaluated with this configuration
locale-formatting96Not evaluated with this configuration
synthetic-hard-negatives540Not 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.

Traced samples for Nightly · v3.2 · 2026-07-12

My phone shows 4G but mobile data does not work after the iOS update.

Device troubleshooting

Passjudge 0.88

How do I get a copy of the personal data Harbor holds about me?

Data privacy requests

Passjudge 0.86

What is the late fee if I pay my bill after the due date?

Billing & payments

Passjudge 0.91

What is the excess on a theft insurance claim?

Returns & warranty

Passjudge 0.90

Can I return a phone I bought online if I opened the box?

Returns & warranty

Passjudge 0.92

My bill says 41,20 €. Is that forty-one euros or four thousand?

Billing & payments

Passjudge 0.89

My eSIM QR code says it has already been used. What should I do?

eSIM & activation

Passjudge 0.86

Can I return my phone after 20 days if I don't like it?

Returns & warranty

FailIncorrectjudge 0.11

Is SIM swap protection turned on by default?

Account & security

Passjudge 0.92

What happens if I go over my fair-use limit on Unlimited at home?

Plans & upgrades

FailMissingjudge 0.07