eval-harness

metric

judge_qwen_s_unvalidated

better
higher
kind
quality: measured against a reference or by a judge
decides
no goal
weighed by
no goal
recorded on
cnn_dailymail_200

01what it measures, and how

Judge qwen_s's mean score for the output, on the scorer's scale, not validated against human ratings.

A project explains its own metric at length with METRIC_DESCRIPTIONS = {"judge_qwen_s_unvalidated": {"summary": "…", "details": "…"}}.

02on cnn_dailymail_200

judge judge-summ-pointwise-v1cnn_deepseek_m_n200_v1 · judged/judge-summ-pointwise-v1: 4.884cnn_bart_l_n200_v1 · judged/judge-summ-pointwise-v1: 4.0614.061better · 4.884

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean judge_qwen_s_unvalidateditemsscored against
cnn_deepseek_m_n200_v1 · judged/judge-summ-pointwise-v14.884200judge judge-summ-pointwise-v1
cnn_bart_l_n200_v1 · judged/judge-summ-pointwise-v14.061200judge judge-summ-pointwise-v1