Study finds LLM judges may misinterpret psychological depth in LLM outputs
This matters for anyone using LLMs to generate content, as it shows that relying solely on LLM judges can lead to incorrect conclusions about the quality or depth of the output.