index / anweat/jev-websearch-eval
jev-websearch-eval
Independent offline evaluation of the Bocha Jev decision model in a web-search evidence pipeline (vs. rule baseline and Laya): tasks, metadata-only snapshots, LLM-draft labels, judge outputs, reports
Research and EvalsRankingListedJavaScript
Repository
Stars0
LanguageJavaScript
LicenseMIT
Created2026-10-06
Last pushed2026-10-06
Officialno
Judgment
Genuine Jev project0.95
gate 0.5
Uses Jev at runtime0.56
Is a meta list0.04
Is a reimplementation0.07
Quality scales
Substance1.90 / 3
gate 0.5
Docs2.97 / 3
Novelty2.65 / 3
Composite score0.80
Category distribution
Research and Evals1.00
Integrations0.00
SDKs and Clients0.00
Applications0.00
Learning0.00
Games and Simulation0.00
Agent and Dev Tooling0.00
Other0.00
Assigned categoryResearch and Evals
Category probability1.00
Category confidence1.00
PatternRanking
Pattern confidence0.67
Curation
StatusListed
Reasongate passed
README pickno
Overridenone
Sources
- search:jev in:name,description,readme created:2026-10-06
Provenance
Modeljev-1.13.0
Question setv2
Judged at2026-10-06 12:43 UTC