← polyphony · eval · frankenstein · ci
polyphony · gemini-2.5-flash
✗ failPass rate
67%
Cases
6 (2 failed)
Mode
eval
Suite
frankenstein
Env
ci
Started
2026-07-20 09:42 UTC
Duration
—
Cost
—
Tokens in/out
— / —
Commit
—
Branch
—
Source
—
Labels
| legacy_env | ci-frankenstein |
Compared with the previous run
Baseline 2026-07-14 02:33 UTC · 67% → this run 67%
✓ newly passing · 1
- eval:ingestion
cases added / removed
- + eval:attribution
- + eval:continuity
- + eval:prose
Cases (6, 2 failed, failures first)
eval:attributionscore 0.33✗ fail
Output
accuracy=0.3333 chance=0.3333 n=3
eval:prosescore 0.00✗ fail
Output
words=0
⚖︎ Judge rationale
empty scene (terminal status=failed: generation failed, blocked, or quota-exhausted)
eval:continuityscore 1.00✓ pass
Output
detection_recall=1.0 fpr=1.0
eval:extractionscore 0.59✓ pass
Output
F1=0.5882 P=0.5 R=0.7143 predicted=['Mrs. Sarrow', 'Isolde', 'Uncle Thomas', 'Hollis Crane', 'Vandel', 'Cosima', 'Alina', 'Fenwick', 'The Creature', 'The Stranger']
eval:ingestionscore 0.63✓ pass
Output
grounded=5/8 avg_chunks=5.62
eval:retrievalscore 0.85✓ pass
Output
precision@k=0.8519 mrr=0.8704