Skip to content

run_2026-08-24_nightly

Nightly · v3.3 · 2026-08-24

Claude Sonnet 5.5 · bge-m3 · hybrid (BM25 + dense, RRF) · support-prompt v6 on support-golden-v4 v4.6

CompletedNightlyIn productionsupport-assistant v3.3v2 · 512 tokens / 64 overlap
Compare

Stratified slice of support-golden-v4 (417 items).

Pass rate

81.3%

0.0 pts(no change)vs previous nightly339 of 417 samples passed

Faithfulness

0.86

−0.01(regression)vs previous nightly

Answer correctness

0.84

+0.01(improvement)vs previous nightly

Context precision

0.73

0.00(no change)vs previous nightly

Context recall

0.79

0.00(no change)vs previous nightly

Latency p95

3.7s

+0.1s(regression)vs previous nightly

Failures by category

78 failed samples · largest share: unsupported

Share of failed samples by category, compared with Nightly · v3.3 · 2026-08-23
CategorySharevs reference
Unsupported claims25.6%−0.3 pts
Incorrect factual answers20.5%−0.3 pts
Missing context21.8%+1.0 pts
Retrieval mismatch20.5%−0.3 pts
Instruction-following failures9.0%−0.1 pts
Other2.6%0.0 pts

Run details

Dataset
support-golden-v4 v4.6 · 417 items
Generator
Claude Sonnet 5.5
Retriever
bge-m3 · hybrid (BM25 + dense, RRF)
Prompt
support-prompt v6
Judge
LLM judge · Claude Opus 5.5 · rubric v3
Started
Aug 24, 02:00 UTC
Duration
42m 56s
Triggered by
Scheduler
Tokens / request
3,980 in · 221 out
  • nightly
  • golden-slice

All metrics

Against Nightly · v3.3 · 2026-08-23

All metrics for Nightly · v3.3 · 2026-08-24
MetricThis runReferenceChange
Pass rate81.3%81.3%0.0 pts
Faithfulness0.860.87−0.01
Answer correctness0.840.83+0.01
Instruction adherence0.890.890.00
Context precision0.730.730.00
Context recall0.790.790.00
Recall@50.830.820.00
MRR0.690.70−0.01
NDCG@100.760.76−0.01
Latency p501.8s1.9s−0.1s
Latency p953.7s3.7s+0.1s
Cost per request$0.0045$0.0046−0.0001

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 v3.3 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
81.3%
0.860.790.833.7sThis run
billing-edge-cases312
76.0%
0.830.770.813.9sRun · Aug 20
roaming-and-travel188
75.5%
0.840.740.783.8smatrix
device-troubleshooting426
78.9%
0.850.780.824.4smatrix
account-security154
82.5%
0.880.800.843.6smatrix
locale-formatting96
75.0%
0.820.750.794.0smatrix
synthetic-hard-negatives540
70.7%
0.820.680.723.9smatrix

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 · v3.3 · 2026-08-24

Why was I charged twice for my September bill?

Billing & payments

Passjudge 0.90

If I upgrade my plan mid-month, how is my next bill calculated?

Plans & upgrades

Passjudge 0.97

How much data can I use in the EU on the Unlimited Plus plan?

Roaming & travel

Passjudge 0.87

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

eSIM & activation

Passjudge 0.93

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

Billing & payments

Passjudge 0.87

What is the excess on a theft insurance claim?

Returns & warranty

Passjudge 0.88

Is SIM swap protection turned on by default?

Account & security

Passjudge 0.94

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

Device troubleshooting

Passjudge 0.87

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

Returns & warranty

Passjudge 0.94

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

Data privacy requests

Passjudge 0.93