Causal Intervention Framework for Variational Auto Encoder Mechanistic Interpretability
Citations
0
Open access
No
Source
crossref
OpenAlex
Not enriched
DOI
10.21203/rs.3.rs-6599791/v1
Abstract
Abstract Mechanistic interpretability of deep learning models has emerged as a crucial research direction for understanding the functioning of neural networks. While significant progress has been made in interpreting discriminative models like transformers, understanding generative models such as Variational Autoencoders (VAEs) remains challenging. This paper introduces a comprehensive causal intervention framework for mechanistic interpretability of VAEs. We develop techniques to identify and analyze "circuit motifs" in VAEs, examining how semantic factors are encoded, processed, and disentangled through the network layers. Our approach uses targeted interventions at different levels: input manipulations, latent space perturbations, activation patching, and causal mediation analysis. We apply our framework to both synthetic datasets with known causal relationships and standard disentanglement benchmarks. Results show that our interventions can successfully isolate functional circuits, map computational graphs to causal graphs of semantic factors, and distinguish between polysemantic and monosemantic units. Furthermore, we introduce metrics for causal effect strength, intervention specificity, and circuit modularity that quantify the interpretability of VAE components. Experimental results demonstrate clear differences between VAE variants, with FactorVAE achieving higher disentanglement scores (0.084) and effect strengths (mean 4.59) compared to standard VAE (0.064, 3.99) and \(\beta\)-VAE (0.051, 3.43). Our framework advances the mechanistic understanding of generative models and provides tools for more transparent and controllable VAE architectures.
Collections
Add to collection
Paper intelligence
Analysis has not been completed yet.
No graph connections yet.
Sync citations or add papers to shared collections to build this network.
Knowledge graph
Citation network
Explore references, papers that cite this work and related papers in your Codex library.
References
0No references have been linked yet.
Cited by
0No saved paper is currently linked as citing this work.
Related papers
0Add papers to shared collections or enrich their topics to find related work.
Research workspace
Attach the paper PDF, extract its text, classify its contents and create semantic embeddings.
Paper resources
No PDF assets have been attached to this paper yet.