Tracer
← polyphony · eval · dracula · ci

polyphony · claude-opus-5-5

✓ pass
Pass rate
100%
Cases
5
Mode
eval
Suite
dracula
Env
ci
Started
2026-10-24 06:14 UTC
Duration
—
Cost
—
Tokens in/out
— / —
Commit
Branch
main
Source

Metrics

attribution.accuracy0.600 (was 0.467)
attribution.chance0.333
attribution.score0.600 (was 0.467)
continuity.detection_recall1
continuity.false_positive_rate0.500 (was 1)
continuity.score1
extraction.f10.609 (was 0.636)
extraction.precision0.438 (was 0.467)
extraction.recall1
extraction.score0.609 (was 0.636)
ingestion.score0.647 (was 0.706)
retrieval.mrr0.955
retrieval.precision_at_k0.818
retrieval.score0.818

Labels

cadenceon-demand
experimentw15-post-2
harness_shaececd1595c9bfb5b766355795d60833b965ec6a6
judge_providergroq
judge_self_gradedFalse
repeat1
triggerworkflow_dispatch

Compared with the previous run

Baseline 2026-10-24 05:58 UTC · 100% → this run 100%

± score moved · 2
  • eval:ingestion: 0.71 → 0.65
  • eval:attribution: 0.47 → 0.60

Cases (5)

gated 4/4 passed

eval:continuityscore 1.00✓ pass
score=1precision=1n_findings=3n_injected=2quote_verbatim=1detection_recall=1n_control_findings=1false_positive_rate=0.500
Output
detection_recall=1.0 fpr=0.5
eval:extractionscore 0.61✓ pass
f1=0.609score=0.609recall=1windows=3precision=0.438f1_selected=0.706recall_selected=0.857precision_selected=0.600principal_fields_rate=1principal_mentions_min=19
Output
F1=0.6087 P=0.4375 R=1.0 predicted=['Cora Ellis', 'Percy Alderton', 'Dr. Elias Ward', 'Van Verhoeven', 'Aldous Kerr', 'Mrs. Ellis', 'Peter Falk', 'Count Vasska', 'Nora Vance', 'Cassidy P. Boone', 'Renwick', 'Mina Vance', 'Sister Agatha', 'R. M. Renwick', 'Vanderpool', 'old Mr. Swales']
eval:ingestionscore 0.65✓ pass
score=0.647grounded=11avg_chunks=11.59n_characters=17
Output
grounded=11/17 avg_chunks=11.59
eval:retrievalscore 0.82✓ pass
mrr=0.955score=0.818n_queries=11precision_at_k=0.818
Output
precision@k=0.8182 mrr=0.9545

reported 1 reported

eval:attribution0.60 · reported
n=30empty=0score=0.600chance=0.333correct=18accuracy=0.600mean_gap=0.046generated=30
Output
accuracy=0.6 chance=0.3333 n=30