arm
cnn_gemma_s_n200_v1 · silver
One config on cnn_dailymail_200, scored one way; its repeats averaged per item.
- dataset
- cnn_dailymail_200
- scored against
- silver/cnn_dailymail_200 silver
- goal
- Short news summaries (CNN)
- runs
- 1
- headline
- coverage 0.350
- last run
- 2026-10-25 16:49
01metrics
| metric | mean | items | better |
|---|---|---|---|
| coverage | 0.350 | 200 | higher |
| compression | 0.131 | 200 | neutral |
| concision | 0.494 | 200 | higher |
| cost_usd | $0.000100 | 200 | lower |
| fmt_bullets | 0.000 | 200 | neutral |
| fmt_label | 0.000 | 200 | neutral |
| fmt_markdown | 0.000 | 200 | neutral |
| fmt_narration | 0.010 | 200 | neutral |
| fmt_overlong | 0.000 | 200 | neutral |
| fmt_paragraphs | 0.000 | 200 | neutral |
| grounding | 0.259 | 200 | higher |
| latency_ms | 6956.2 | 200 | lower |
| length_vs_reference | 0.721 | 200 | neutral |
| reasoning_tokens | 0.000 | 200 | neutral |
| reference_words | 74.5 | 200 | neutral |
| rouge1 | 0.470 | 200 | higher |
| rougeLsum | 0.406 | 200 | higher |
| summary_words | 52.5 | 200 | neutral |
| tokens_in | 750.7 | 200 | lower |
| tokens_out | 69.6 | 200 | lower |
| truncated | 0.000 | 200 | neutral |
02cost and speed
- cost per item
- $0.000100 · 820 tok
- speed per item
- p50 5,933 · p95 12,255 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 |
|---|---|---|---|---|
cnn_gemma_s_n200_v1_20260926T073908Z | 2026-10-25 16:49 | 200 | $0.020016 | rescore |
04config
config cnn_gemma_s_n200_v1 · adapter examples/summarization-cnn-dailymail/adapter.py · credentials are never shown
- family
gemma- max_tokens
1200- model
eval-gemma-3-27b- price_tier
small- prompt_file
prompt.txt- provider
litellm- reasoning
{"enabled": false}- temperature
0.0- usd_per_mtok_in
0.07- usd_per_mtok_out
0.15
05conditions
- data.items_sha256
257feda902391b934b8cde940c15e91201096dbe96ec5a81337e9da007ce3ec7- deployment.model
eval-gemma-3-27b- deployment.source
adapter- deployment.upstream
openrouter/google/gemma-3-27b-it- fingerprint.version
5- library.numpy
2.4.6- library.openai
3.19.2- library.rouge-score
0.1.2- machine.cpu_count
12- machine.platform
Darwin/x86_64- machine.processor
i386- machine.python
3.11.8- machine.release
24.6.0- reference.sha256
b7cd73d14c009e1ee5559a332cfc37143e24b4debcc75d1fdda9235a85b26f3d- scorer.id
examples/summarization-cnn-dailymail/adapter.py- scorer.sha256
3826aad41bf91e498b57f98536babef281452cd09b4eb793e20dec53fd9bc4f8
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
| arm | scored against | headline | mean |
|---|---|---|---|
| cnn_gemma_s_n200_v1 | gold/cnn_dailymail_200 gold | coverage | 0.312 |
| cnn_gemma_s_n200_v1 · judged/judge-summ-pick-v1 | judge judge-summ-pick-v1 | judge_won | 0.002 |
| cnn_gemma_s_n200_v1 · judged/judge-summ-v1 | judge judge-summ-v1 | judge_score | 4.801 |
| cnn_gemma_s_n200_v1 · judged/judge-summ-v2 | judge judge-summ-v2 | judge_score | 4.429 |
07rounds
- Round r1 of Short news summaries (CNN), closed: one of its arms.