Learning systems that preserve privacy often inject noise into hierarchical visual representations; a central challenge is to \emph{model} how such perturbations align with a declared privacy budget in a way that is interpretable and applicable across vision backbones and vision--language models (VLMs). We propose \emph{Bodhi VLM}, a \emph{privacy-alignment modeling} framework for \emph{hierarchical neural representations}: it (1) links sensitive concepts to layer-wise grouping via NCP and MDAV-based clustering; (2) locates sensitive feature regions using bottom-up (BUA) and top-down (TDA) strategies over multi-scale representations (e.g., feature pyramids or vision-encoder layers); and (3) uses an Expectation-Maximization Privacy Assessment (EMPA) module to produce an interpretable \emph{budget-alignment signal} by comparing the fitted sensitive-feature distribution to an evaluator-specified reference (e.g., Laplace or Gaussian with scale $c/ε$). The output is reference-relative and is \emph{not} a formal differential-privacy estimator. We formalize BUA/TDA over hierarchical feature structures and validate the framework on object detectors (YOLO, PPDPTS, DETR) and on the \emph{visual encoders} of VLMs (CLIP, LLaVA, BLIP). BUA and TDA yield comparable deviation trends; EMPA provides a stable alignment signal under the reported setups. We compare with generic discrepancy baselines (Chi-square, K-L, MMD) and with task-relevant baselines (MomentReg, NoiseMLE, Wass-1). Results are reported as mean$\pm$std over multiple seeds with confidence intervals in the supplementary materials. This work contributes a learnable, interpretable modeling perspective for privacy-aligned hierarchical representations rather than a post hoc audit only. Source code: \href{https://github.com/mabo1215/bodhi-vlm.git}{Bodhi-VLM GitHub repository}

翻译：在保护隐私的学习系统中，噪声常被注入到分层视觉表征中；一个核心挑战是如何以可解释且适用于各类视觉骨干网络和视觉-语言模型（VLM）的方式，**建模**此类扰动与声明的隐私预算之间的**对齐关系**。我们提出**Bodhi VLM**，一个面向**分层神经表征**的**隐私对齐建模**框架：它（1）通过基于NCP和MDAV的聚类方法，将敏感概念与逐层分组相关联；（2）利用自底向上（BUA）和自顶向下（TDA）策略在多尺度表征（例如特征金字塔或视觉编码器层）上定位敏感特征区域；（3）采用期望最大化隐私评估（EMPA）模块，通过将拟合的敏感特征分布与评估者指定的参考分布（例如尺度为 $c/ε$ 的拉普拉斯分布或高斯分布）进行比较，生成可解释的**预算对齐信号**。该输出是相对于参考的，并**非**一个正式的差分隐私估计量。我们形式化了在分层特征结构上的BUA/TDA方法，并在目标检测器（YOLO、PPDPTS、DETR）和VLM的**视觉编码器**（CLIP、LLaVA、BLIP）上验证了该框架。BUA和TDA产生了可比较的偏差趋势；EMPA在所述设置下提供了稳定的对齐信号。我们与通用差异基线（卡方检验、K-L散度、MMD）以及任务相关基线（MomentReg、NoiseMLE、Wass-1）进行了比较。结果以多次随机种子下的均值±标准差形式报告，置信区间见补充材料。本工作贡献了一种可学习的、可解释的隐私对齐分层表征建模视角，而非仅提供事后审计。源代码：\href{https://github.com/mabo1215/bodhi-vlm.git}{Bodhi-VLM GitHub 仓库}