Who Judges the Judge? The Rise of Meta-Evaluation in AI
中文摘要
AI被用于评估AI，但评估者同样存在偏差，引发了关于元评估可靠性的挑战。
English Summary
Using AI to evaluate AI reveals that evaluators are also biased, raising critical questions about the reliability of meta-evaluation.
Original Excerpt
We taught AI to evaluate AI. Then we discovered the evaluator was biased too. Now what? Continue reading on Learning Data »