The Quantum Sieve Tracer: A Hybrid Framework for Layer-Wise Activation Tracing in Large Language Models

TL;DR

This paper presents the Quantum Sieve Tracer, a hybrid quantum-classical framework for detailed layer-wise activation analysis in large language models, revealing distinct architectural functions and mechanisms.

Contribution

It introduces a novel hybrid quantum-classical method for mechanistic interpretability, enabling high-resolution analysis of attention mechanisms in LLMs.

Findings

Qwen's layer 7 functions as a Recall Hub.

Llama's layer 9 acts as an Interference Suppression circuit.

Quantum kernels distinguish constructive and reductive attention mechanisms.

Abstract

Mechanistic interpretability aims to reverse-engineer the internal computations of Large Language Models (LLMs), yet separating sparse semantic signals from high-dimensional polysemantic noise remains a significant challenge. This paper introduces the Quantum Sieve Tracer, a hybrid quantum-classical framework designed to characterize factual recall circuits. We implement a modular pipeline that first localizes critical layers using classical causal tracing, then maps specific attention head activations into an exponentially large quantum Hilbert space. Using open-weight models (Meta Llama-3.2-1B and Alibaba Qwen2.5-1.5B-Instruct), we perform a two-stage analysis that reveals a fundamental architectural divergence. While Qwen's layer 7 circuit functions as a classic Recall Hub, we discover that Llama's layer 9 acts as an Interference Suppression circuit, where ablating the identified heads paradoxically improves factual recall. Our results demonstrate that quantum kernels can distinguish between these constructive (recall) and reductive (suppression) mechanisms, offering a high-resolution tool for analyzing the fine-grained topology of attention.