OPEN-SOURCE PROJECT

jevcal

abhixhek · Evaluation & Observability

A toolkit for evaluating Jev probabilities on labeled data, selecting confidence thresholds and checking model drift.

JEV · DECISION POINT

What Jev does here

Runs fixed questions and measures accuracy, calibration, coverage and escalation rates.

WHY IT HELPS

Connects threshold selection and model-change checks to reports and CI.

SOURCE NOTE

The README demo table uses a simulator, not measured Jev results; production thresholds need workload-specific validation. Not run here.

REVIEWED SOURCE LINKS