GLIDER: Grading LLM Interactions and Decisions using Explainable Ranking

Darshan Deshpande, Selvan Sunitha Ravi, Sky CH-Wang +3

The LLM-as-judge paradigm is increasingly being adopted for automated evaluation of model outputs. While LLM judges have shown promise on constrained evaluation tasks, closed sourc…