CRUXEval-Output-CoT

CRUXEval-Output-CoT enhances the standard CRUXEval output prediction task by incorporating chain-of-thought reasoning methodology. This benchmark evaluates Large Language Models' ability to predict program outputs while showing their reasoning process step-by-step. CRUXEval-Output-CoT tests both code execution understanding and the capacity to articulate logical reasoning in code analysis scenarios.