1 paper
Zhuohang Li, Xiaowei Li, Chengyu Huang +11
The alignment of large language models (LLMs) with human values increasingly relies on using other LLMs as automated judges, or ``autoraters''. However, their reliability is limite…