/ github.com / gke-labs / eval-runner
← Back to Overview
0.69 [0.44, 0.86] Δ +0.12 [−0.19, +0.41]
docs · 2 evals · at commit 2a0e7b7

docs/readme_comprehension

1.00 [0.68, 1.00]
Δ 0.00 [−0.32, +0.32]
0.00 0.50 1.00 0a3e6b0 2a0e7b7
first: 0a3e6b0 last: 2a0e7b7 newest-changed: 2a0e7b7
Commit Time Model Samples Score (95% CI) Action
2a0e7b7 55m google/gemini-3.5-flash-lite 8 1.00 [0.68, 1.00] view samples →
0a3e6b0 2d google/gemini-3.5-flash-lite 8 1.00 [0.68, 1.00] view samples →

docs/readme_comprehension_blind

0.38 [0.14, 0.69]
Δ +0.25 [−0.17, +0.59]
0.00 0.50 1.00 0a3e6b0 2a0e7b7
first: 0a3e6b0 last: 2a0e7b7 newest-changed: 2a0e7b7
Commit Time Model Samples Score (95% CI) Action
2a0e7b7 55m google/gemini-3.5-flash-lite 8 0.38 [0.14, 0.69] view samples →
0a3e6b0 2d google/gemini-3.5-flash-lite 8 0.12 [0.02, 0.47] view samples →
Score at main is the pooled Wilson success rate over the newest completed baseline observations at the current commit. An observation is one completed task result on one Run.