Lab

LLM-as-Judge — Evaluating AI Responses with AI

Analysis of the LLM-as-Judge pattern for evaluating AI response quality, featuring multidimensional metric design, reliability verification, and strategies for position and verbosity bias.