/FormatEval

[Preprint' 24] LLMs Are Biased Towards Output Formats! Systematically Evaluating and Mitigating Output Format Bias of LLMs

Primary LanguagePython

Stargazers