Mechanistic Interpretability in the Presence of Architectural Obfuscation

TL;DR

This paper investigates how architectural obfuscation techniques affect the interpretability of large language models, showing that while they preserve overall function, they hinder detailed mechanistic understanding.

Contribution

It provides a systematic analysis of obfuscation's impact on interpretability, revealing that it degrades circuit-level insights without affecting model performance.

Findings

Obfuscation alters attention head activation patterns.

Layer-wise computational graphs remain intact.

Fine-grained interpretability is significantly impaired.

Abstract

Architectural obfuscation - e.g., permuting hidden-state tensors, linearly transforming embedding tables, or remapping tokens - has recently gained traction as a lightweight substitute for heavyweight cryptography in privacy-preserving large-language-model (LLM) inference. While recent work has shown that these techniques can be broken under dedicated reconstruction attacks, their impact on mechanistic interpretability has not been systematically studied. In particular, it remains unclear whether scrambling a network's internal representations truly thwarts efforts to understand how the model works, or simply relocates the same circuits to an unfamiliar coordinate system. We address this gap by analyzing a GPT-2-small model trained from scratch with a representative obfuscation map. Assuming the obfuscation map is private and the original basis is hidden (mirroring an honest-but-curious server), we apply logit-lens attribution, causal path-patching, and attention-head ablation to locate and manipulate known circuits. Our findings reveal that obfuscation dramatically alters activation patterns within attention heads yet preserves the layer-wise computational graph. This disconnect hampers reverse-engineering of user prompts: causal traces lose their alignment with baseline semantics, and token-level logit attributions become too noisy to reconstruct. At the same time, feed-forward and residual pathways remain functionally intact, suggesting that obfuscation degrades fine-grained interpretability without compromising top-level task performance. These results establish quantitative evidence that architectural obfuscation can simultaneously (i) retain global model behaviour and (ii) impede mechanistic analyses of user-specific content. By mapping where interpretability breaks down, our study provides guidance for future privacy defences and for robustness-aware interpretability tooling.