arm
triage_gpt_61_sol_v1 · repeats ×3
One config on orrery_triage_v1_labelled, scored one way; its repeats averaged per item.
- dataset
- orrery_triage_v1_labelled
- scored against
- gold/orrery_triage_v1_labelled gold
- goal
- Triage decisions on Orrery's error signals
- runs
- 3
- headline
- false_dismiss 0.088
- last run
- 2026-10-30 14:37
- re-scored
- the goals read its re-score, runs-rescored/5f2c057cdaa5/runs-repeats/orrery_triage_v1_labelled/triage_gpt_61_sol_v1
Inspect itAgainst the baselineWith the 19 arms scored the same way
01metrics
| metric | mean | items | better |
|---|---|---|---|
| false_dismiss | 0.088 | 34 | lower |
| correct | 0.600 | 65 | higher |
| cost_usd | $0.010918 | 65 | lower |
| escalated | 0.185 | 65 | lower |
| false_alarm | 0.742 | 31 | lower |
| latency_ms | 12005.5 | 65 | lower |
| model_correct | 0.641 | 65 | higher |
| model_escalated | 0.062 | 65 | lower |
| model_false_alarm | 0.505 | 31 | lower |
| model_false_dismiss | 0.108 | 34 | lower |
| n_probes | 1.810 | 65 | higher |
| table_miss | 0.000 | 65 | lower |
| tokens_in | 3835.0 | 65 | lower |
| tokens_out | 324.8 | 65 | lower |
02cost and speed
- cost per item
- $0.010918 · 4,160 tok
- speed per item
- p50 11,696 · p95 18,138 ms
Per item, over its items: cost as dollars and tokens (a local model is free but still uses tokens), speed as the median and the slow tail (p95). A goal that names this arm weighs them in its decision.
03runs
| run | made | items | cost | how |
|---|---|---|---|---|
triage_gpt_61_sol_v1_20261030T135822Z_r1 | 2026-10-30 14:11 | 65 | $0.706256 | measured |
triage_gpt_61_sol_v1_20261030T135822Z_r2 | 2026-10-30 14:24 | 65 | $0.709608 | measured |
triage_gpt_61_sol_v1_20261030T135822Z_r3 | 2026-10-30 14:37 | 65 | $0.713158 | measured |
04config
config triage_gpt_61_sol_v1 · adapter examples/triage-fleet/adapter.py · credentials are never shown
- model
eval-gpt-61-sol- provider
gateway- usd_per_mtok_in
2.0- usd_per_mtok_out
10.0
05conditions
- data.items_sha256
d428aca49adc8e762dcbbcad6c9a1dba1ac4e4e3315ff26025af975faba1176a- deployment.model
eval-gpt-61-sol- deployment.source
adapter- fingerprint.version
5- harness.code_sha256
0ab386969208f46c7448153422d5a5ab07833d963869b69fc216a6cd28feb110- harness.version
0.5.0- library.anthropic
2.3.0- library.numpy
2.5.4- library.openai
3.41.0- library.rouge-score
0.1.2- machine.cpu_count
32- machine.platform
Linux/x86_64- machine.python
3.14.7- machine.release
7.2.5-3-omarchy- reference.sha256
f617ad3bb9bb9f5feb8ffe54e0ef1b9b05dfb316fac13f63a925f9d35fb191cd- scorer.id
examples/triage-fleet/adapter.py- scorer.sha256
9e899d6853cb4065ab07f1d44c08120d0a19ad282abdfdc0e95a3e63262cca81
06the same config, scored other ways
scored against silverscored by a judgescored against a label set, not a reference setunmarked rows are scored against gold