论文专业

Faithfulness

在可解释 AI 评估中，指 AI 生成的解释与模型真实推理过程的一致程度，区别于看起来合理但实际并不忠实于模型的'事后合理化'解释。

English Definition

"In XAI evaluation, the degree to which an AI-generated explanation accurately reflects the model's actual reasoning process, as opposed to being merely plausible or post-hoc rationalized."

常见错误

·不要把 faithfulness 简单翻译成'真实性'，在 XAI 里特指解释与模型内部推理的一致性。

真实用例 · 3 条

"We conducted a three-phase evaluation following a multi-level validation framework: a functional evaluation of faithfulness, a clinician evaluation of workflow suitability, and a patient evaluation of perceived understanding and trust."

论文Pipeline extracted — review needed·Faithfulness 作为多层次评估中第一阶段（功能评估）的评估对象

"How well do MAP-X explanations demonstrate functional faithfulness by aligning with an expert-defined structure, the model's internal signal, and the ground-truth data source?"

论文Pipeline extracted — review needed·研究问题 RQ1 直接探讨功能忠实度

"MAP-X produced clinically relevant explanations with reasonable faithfulness and usability."

论文Pipeline extracted — review needed·Faithfulness 作为系统整体质量指标之一

由 pipeline 自动采集，待人工 review