human verification

New framework AURA refines LLM-as-a-Judge auditing

Researchers have introduced AURA, a novel framework designed to improve the auditing of large language models (LLMs) when they are used as judges in evaluations. AURA addresses the challenge that LLM judges can be biase…