LLM-as-a-Judge systems exhibit significant biases in specific tasks despite strong overall performance, as measured by the new CALM quantification framework.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge
LLM-as-a-Judge systems exhibit significant biases in specific tasks despite strong overall performance, as measured by the new CALM quantification framework.