Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.03530 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914443930107904 |
|---|---|
| author | Roy, Dip Misra, Rajiv Singh, Sanjay Kumar Roy, Anisha |
| author_facet | Roy, Dip Misra, Rajiv Singh, Sanjay Kumar Roy, Anisha |
| contents | Understanding how generative models represent and transform data is a foundational problem in deep learning interpretability. While mechanistic interpretability of discriminative architectures has yielded substantial insights, relatively little work has addressed variational autoencoders (VAEs). This paper presents the first general-purpose multilevel causal intervention framework for mechanistic interpretability of VAEs. The framework comprises four manipulation types: input manipulation, latent-space perturbation, activation patching, and causal mediation analysis. We also define three new quantitative metrics capturing properties not measured by existing disentanglement metrics alone: Causal Effect Strength (CES), intervention specificity, and circuit modularity. We conduct the largest empirical study to date of VAE causal mechanisms across six architectures (standard VAE, beta-VAE, FactorVAE, beta-TC-VAE, DIP-VAE-II, and VQ-VAE) and five benchmarks (dSprites, 3DShapes, MPI3D, CelebA, and SmallNORB), with three seeds per configuration, totaling 90 independent training runs. Our results reveal several findings: (i) a consistent within-dataset negative correlation between CES and DCI disentanglement (the CES-DCI trade-off); (ii) that the KL reweighting mechanism of beta-VAE induces a capacity bottleneck when generative factors approach latent dimensionality, degrading disentanglement on complex datasets; (iii) that no single VAE architecture dominates across all five datasets, with optimal choice depending on dataset structure; and (iv) that CES-based metrics applied to discrete latent spaces (VQ-VAE) yield near-zero values, revealing a critical limitation of continuous-intervention methods for discrete representations. These results provide both a theoretical foundation and comprehensive empirical evaluation for mechanistic interpretability of generative models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_03530 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Multi-Level Causal Intervention Framework for Mechanistic Interpretability in Variational Autoencoders Roy, Dip Misra, Rajiv Singh, Sanjay Kumar Roy, Anisha Machine Learning Understanding how generative models represent and transform data is a foundational problem in deep learning interpretability. While mechanistic interpretability of discriminative architectures has yielded substantial insights, relatively little work has addressed variational autoencoders (VAEs). This paper presents the first general-purpose multilevel causal intervention framework for mechanistic interpretability of VAEs. The framework comprises four manipulation types: input manipulation, latent-space perturbation, activation patching, and causal mediation analysis. We also define three new quantitative metrics capturing properties not measured by existing disentanglement metrics alone: Causal Effect Strength (CES), intervention specificity, and circuit modularity. We conduct the largest empirical study to date of VAE causal mechanisms across six architectures (standard VAE, beta-VAE, FactorVAE, beta-TC-VAE, DIP-VAE-II, and VQ-VAE) and five benchmarks (dSprites, 3DShapes, MPI3D, CelebA, and SmallNORB), with three seeds per configuration, totaling 90 independent training runs. Our results reveal several findings: (i) a consistent within-dataset negative correlation between CES and DCI disentanglement (the CES-DCI trade-off); (ii) that the KL reweighting mechanism of beta-VAE induces a capacity bottleneck when generative factors approach latent dimensionality, degrading disentanglement on complex datasets; (iii) that no single VAE architecture dominates across all five datasets, with optimal choice depending on dataset structure; and (iv) that CES-based metrics applied to discrete latent spaces (VQ-VAE) yield near-zero values, revealing a critical limitation of continuous-intervention methods for discrete representations. These results provide both a theoretical foundation and comprehensive empirical evaluation for mechanistic interpretability of generative models. |
| title | A Multi-Level Causal Intervention Framework for Mechanistic Interpretability in Variational Autoencoders |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2505.03530 |