ConsistencyChecker: Tree-based Evaluation of LLM Generalization Capabilities

Zhaochen Hong, Haofei Yu, Jiaxuan You

Evaluating consistency in large language models (LLMs) is crucial for ensuring reliability, particularly in complex, multi-step interactions between humans and LLMs. Traditional se…