eval-harness

arm

cnn_lead3_v1

One config on cnn_dailymail_20, scored one way; its repeats averaged per item.

dataset
cnn_dailymail_20
scored against
gold/cnn_dailymail_20 gold
goal
Short news summaries (CNN)
runs
1
headline
coverage 0.262
last run
2026-09-26 15:33

Inspect itWith the 2 arms scored the same way

01metrics

metricmeanitemsbetter
coverage0.26220higher
compression0.17320neutral
concision0.22520higher
cost_usd$0.00000020lower
fmt_bullets0.00020neutral
fmt_label0.00020neutral
fmt_markdown0.00020neutral
fmt_narration0.00020neutral
fmt_overlong0.05020neutral
fmt_paragraphs0.00020neutral
grounding1.00020higher
latency_ms0.07120lower
length_vs_reference1.71620neutral
reference_words36.720neutral
rouge10.31220higher
rougeLsum0.28020higher
summary_words61.420neutral

02cost and speed

cost per item
$0.000000
speed per item
p50 0.069 · p95 0.13 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

runmadeitemscosthow
cnn_lead3_v1_20260926T153351Z2026-09-26 15:3320$0.000000measured

04config

config cnn_lead3_v1 · adapter examples/summarization-cnn-dailymail/adapter.py · credentials are never shown

family
lead
price_tier
free
provider
lead_k
sentences
3

05conditions

data.items_sha256
ea2e08a2214a375f00db936da844b4592fd2cd3f93d1f1460735231b7bb83f02
deployment.source
adapter
fingerprint.version
5
library.numpy
1.26.4
library.openai
3.19.2
library.rouge-score
0.1.2
library.torch
2.2.2
library.transformers
4.55.4
machine.cpu_count
12
machine.label
mac-dev
machine.platform
Darwin/x86_64
machine.processor
i386
machine.python
3.11.8
machine.release
24.6.0
reference.sha256
7193289e59b4438f7191ddc149929d6ca637080154b9fca52d5ba3a7259c57a5
scorer.id
examples/summarization-cnn-dailymail/adapter.py
scorer.sha256
51a7b06b1a2eec722a3bd024b7ed832746d06ef44683dd0e9fa8abc5c18d5618