Mechanistic Interpretability and Causal Feature Steering of Neural Quantum States via Sparse Autoencoders

• URL: http://arxiv.org/abs/2607.01336v1
• Date: Wed, 01 Jul 2026 18:00:11 GMT
• Title: Mechanistic Interpretability and Causal Feature Steering of Neural Quantum States via Sparse Autoencoders
• Abstract summary: We present a systematic approach to analyze the internal activations of Neural Quantum States (NQS) using sparse autoencoders.<n>We find that features strongly correlate with physical observables such as order parameters, staggered magnetization, and half-chain correlators.<n>We further establish that such features causally affect the corresponding observables predicted by NQS, by showing that targeted, post-training intervention on a textitsingle feature smoothly steers the corresponding observable.
• Score: 0.14259046373656994
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Neural Quantum States (NQS) are a remarkably expressive class of variational ansätze for quantum many-body wavefunctions, yet little is understood about their internal mechanisms: trained on variational objectives alone, how do NQS accurately capture physical observables that they have never been explicitly optimized for? In this work, we present a systematic approach to analyze the internal activations of NQS using sparse autoencoders. We extract features from the residual stream and demonstrate that these features strongly correlate with physical observables such as order parameters, staggered magnetization, and half-chain correlators, across both ground state representation and real-time dynamics. Remarkably, the discovery of these features is entirely unsupervised, with no physical labels provided. We further establish that such features causally affect the corresponding observables predicted by NQS, by showing that targeted, post-training intervention on a \textit{single} feature smoothly and monotonically steers the corresponding observable, while leaving the variational energy nearly unchanged. These results demonstrate that NQS are not merely functional approximators, but encode rich, interpretable internal representations of physical information. Our approach provides both a diagnostic and an intervention tool for NQS, and serves as a foundation for using mechanistic interpretability towards more reliable, transparent NQS.

Related papers

• Neural Quantum States in Non-Stabilizer Regimes: Benchmarks with Atomic Nuclei [41.99844472131922]
  We study the performance of neural quantum states in relation to the quantum complexity of the target state.<n>For a fixed number of configurations, we find that states with larger non-stabilizerness are systematically harder to learn.<n>This finding suggests that non-stabilizerness is a primary factor governing the compression and representational efficiency of RBMs in entangled regimes.
  arXiv  Detail & Related papers  (2026-03-30T16:29:42Z)
• Basis dependence of Neural Quantum States for the Transverse Field Ising Model [0.0]
  We investigate the dependence of Neural Quantum States on the choice of the computational basis.<n>Considering a family of rotated Hamiltonians corresponding to the paradigmatic transverse-field Ising model, we discuss the properties of ground states responsible for the dependence of NQS performance.<n>We identify that the basis-dependence of the performance is linked to the convergence properties of a cluster or cumulant expansion of multi-spin operators.
  arXiv  Detail & Related papers  (2025-12-12T15:11:05Z)
• Quantum Dynamics with Time-Dependent Neural Quantum States [41.99844472131922]
  We present proof-of-principle time-dependent neural quantum state (NQS) simulations.<n>NQS leverage the parameterization of the wave function with neural-network architectures.
  arXiv  Detail & Related papers  (2025-09-29T14:50:02Z)
• Importance of Correlations for Neural Quantum States [0.0]
  Neural quantum states (NQS) have emerged as a powerful variational ansatz for representing quantum many-body wave functions.<n>We investigate the role of correlations for NQS-like quantum state representation by employing a correlation-based interpretable neural network architecture.
  arXiv  Detail & Related papers  (2025-08-19T17:55:21Z)
• Allostatic Control of Persistent States in Spiking Neural Networks for perception and computation [79.16635054977068]
  We introduce a novel model for updating perceptual beliefs about the environment by extending the concept of Allostasis to the control of internal representations.<n>In this paper, we focus on an application in numerical cognition, where a bump of activity in an attractor network is used as a spatial numerical representation.
  arXiv  Detail & Related papers  (2025-03-20T12:28:08Z)
• Fourier Neural Operators for Learning Dynamics in Quantum Spin Systems [77.88054335119074]
  We use FNOs to model the evolution of random quantum spin systems. We apply FNOs to a compact set of Hamiltonian observables instead of the entire $2n$ quantum wavefunction.
  arXiv  Detail & Related papers  (2024-09-05T07:18:09Z)
• Partially Observable Stochastic Games with Neural Perception Mechanisms [31.51588071503617]
  We propose the model of neuro-symbolic partially-observable games (NS-POSGs) We focus on a one-sided setting with a partially-informed agent using discrete, data-driven observations and another, fully-informed agent. We present a new method, called one-sided NS-HSVI, for approximate solution of one-sided NS-POSGs.
  arXiv  Detail & Related papers  (2023-10-17T20:25:40Z)
• Neural-network quantum states for ultra-cold Fermi gases [49.725105678823915]
  This work introduces a novel Pfaffian-Jastrow neural-network quantum state that includes backflow transformation based on message-passing architecture. We observe the emergence of strong pairing correlations through the opposite-spin pair distribution functions. Our findings suggest that neural-network quantum states provide a promising strategy for studying ultra-cold Fermi gases.
  arXiv  Detail & Related papers  (2023-05-15T17:46:09Z)
• Absence of operator growth for average equal-time observables in charge-conserved sectors of the Sachdev-Ye-Kitaev model [11.353329565587574]
  Quantum scrambling plays an important role in understanding thermalization in closed quantum systems. We show that scrambling is absent for disorder-averaged expectation values of observables. We develop a cumulant expansion approach to approximate the evolution of equal-time observables.
  arXiv  Detail & Related papers  (2022-10-05T17:47:52Z)
• Symmetric Pruning in Quantum Neural Networks [111.438286016951]
  Quantum neural networks (QNNs) exert the power of modern quantum machines. QNNs with handcraft symmetric ansatzes generally experience better trainability than those with asymmetric ansatzes. We propose the effective quantum neural tangent kernel (EQNTK) to quantify the convergence of QNNs towards the global optima.
  arXiv  Detail & Related papers  (2022-08-30T08:17:55Z)
• Momentum Diminishes the Effect of Spectral Bias in Physics-Informed Neural Networks [72.09574528342732]
  Physics-informed neural network (PINN) algorithms have shown promising results in solving a wide range of problems involving partial differential equations (PDEs) They often fail to converge to desirable solutions when the target function contains high-frequency features, due to a phenomenon known as spectral bias. In the present work, we exploit neural tangent kernels (NTKs) to investigate the training dynamics of PINNs evolving under gradient descent with momentum (SGDM)
  arXiv  Detail & Related papers  (2022-06-29T19:03:10Z)
• Autoregressive neural-network wavefunctions for ab initio quantum chemistry [3.5987961950527287]
  We parameterise the electronic wavefunction with a novel autoregressive neural network (ARN) This allows us to perform electronic structure calculations on molecules with up to 30 spin-orbitals.
  arXiv  Detail & Related papers  (2021-09-26T13:44:41Z)
• Quantum-tailored machine-learning characterization of a superconducting qubit [50.591267188664666]
  We develop an approach to characterize the dynamics of a quantum device and learn device parameters. This approach outperforms physics-agnostic recurrent neural networks trained on numerically generated and experimental data. This demonstration shows how leveraging domain knowledge improves the accuracy and efficiency of this characterization task.
  arXiv  Detail & Related papers  (2021-06-24T15:58:57Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.