Post-hoc explanation depends on how model queries are organized. We propose CUBE, a design-based framework that explains a trained predictive model through balanced low--high probes. Selected variables define factors, designed feature-level combinations define query conditions, and model predictions are summarized as factorial contrasts. CUBE reports main effects and pairwise interactions as controlled readings of average and conditional response changes over a declared design space. Experiments on synthetic and real tabular tasks show that CUBE recovers dominant learned effect structure, clarifies query-efficient identifiability, and supports screening--follow-up refinement.

翻译：事后解释依赖于模型查询的组织方式。我们提出CUBE——一种基于设计的框架，通过平衡低-高探针来解释已训练的预测模型。所选变量定义因子，设计的特征级组合定义查询条件，模型预测被汇总为因子对比。CUBE报告主效应和成对交互作用，作为对声明设计空间内平均响应与条件响应的受控解读。在合成及真实表格数据任务上的实验表明，CUBE能够恢复主导的学习效应结构，阐明查询高效的可识别性，并支持筛选-后续精化流程。