Imibuzo exhumene nomhlahlandlela · Kunzima Izinga

Calibrating LLM Judges Against Humans Quiz

Calibrate LLM judges with shared rubrics, independent human labels, confusion matrices, suitable agreement statistics, and repeated held-out checks.

Izindlela zomhlahlandlela ezihlobeneCalibrating LLM Judges Against Humans
Umbuzo 1 kwe 8

Why is Cohen's kappa preferred over raw percent agreement when calibrating an LLM judge?