Search arXiv⌕ Search

arXiv subjects

Jennifer Lin

Publications and source records attributed to Jennifer Lin.

At least 19 recordsLinked to original sources
arXiv2026

Ryu-Takayanagi area from Virasoro modular data

We show that in holographic 2d CFTs, the entanglement entropies across several choices of global state and subregion can be written in a way that at once has a microscopic interpretation and matches the leading large$-c$ organization of the Ryu-Takayanagi formula. This representation is obtained by applying crossing symmetry to the replica manifolds. From the boundary point of view, each rewritten entropy looks like an algebraic entanglement entropy for the Virasoro algebra restricted to the region, with center labels obtained by coarse-graining heavy primaries of the BCFT on the regulated region into bins labeled by Liouville momenta. At large $c$, a resulting sum over bins is dominated by a saddle, and the $O(c)$ part of the entropy comes from the Cardy density of heavy primaries in the dominant bin. We identify this $O(c)$ part of the entropy with the Ryu-Takayanagi area. Physically, this suggests a concrete statistical origin for the Ryu-Takayanagi area as coming from coarse-grained Virasoro intertwiners across the entangling cut. The result also provides quantitative criteria for the amount of coarse-graining allowed for the consistency of the interpretation.

hep-th↗
arXiv2026

Towards Worst-Case Guarantees with Scale-Aware Interpretability

Neural networks organize information according to the hierarchical, multi-scale structure of natural data. Methods to interpret model internals should be similarly scale-aware, explicitly tracking how features compose across resolutions and guaranteeing bounds on the influence of fine-grained structure that is discarded as irrelevant noise. We posit that the renormalisation framework from physics can meet this need by offering technical tools that can overcome limitations of current methods. Moreover, relevant work from adjacent fields has now matured to a point where scattered research threads can be synthesized into practical, theory-informed tools. To combine these threads in an AI safety context, we propose a unifying research agenda -- \emph{scale-aware interpretability} -- to develop formal machinery and interpretability tools that have robustness and faithfulness properties supported by statistical physics.

hep-th↗
arXiv2026

Instabilities of the Fractionalized Dirac Semimetal in the Kitaev-Kondo Model

We study a honeycomb Kondo lattice model in which Dirac conduction electrons are coupled to a spin-1/2 Kitaev quantum spin liquid. For weak Kondo coupling, the spins fractionalize into Majorana fermions comprising a gapless Dirac mode and three gapped visons. In second order perturbation theory, the Kondo coupling gives rise to local Hubbard repulsions and spin-spin interactions between conduction electrons, as well as a vertex coupling electrons to gapless Majorana fermions. We analyze the resulting low-energy field theory using a perturbative renormalization group (RG) scheme, accounting for additional density-density interactions generated under RG. At criticality, electrons decouple from Majorana fermions but all three electron interactions acquire positive values. An analysis of susceptibility exponents reveals that the fractionalized Fermi liquid becomes unstable towards antiferromagnetic order and that superconductivity is disfavored.

cond-mat.str-el↗
arXiv2025

Feature Identification via the Empirical NTK

We provide evidence that eigenanalysis of the empirical neural tangent kernel (eNTK) can surface feature directions in trained neural networks. Across three increasingly realistic settings -- a 1-layer MLP trained on modular addition, a 1-layer Transformer trained on modular addition and the pretrained language model Gemma-3-270M -- we show that top eigenspaces of the eNTK align with ground-truth or interpretable features. In the modular arithmetic examples, top eNTK eigenspaces align with the Fourier features used by the MLP and the Fourier features at seed-dependent frequencies used by the Transformer to implement known ground-truth algorithms. Moreover, the alignment of the relevant subspaces evolves over training, with its first derivative peaking near the onset of grokking. For Gemma-3-270M, we compute top eNTK eigendirections on a dataset of TinyStories context windows and check their alignment with an automatically-generated set of parts-of-speech and other grammatical feature directions. We find that the alignment of eNTK eigendirections with grammar features outperforms a same-budget baseline of PCA on model activations. These results suggest that eNTK eigenanalysis may provide a new handle towards identifying features in trained models for mechanistic interpretability.

cs.LG↗
arXiv2023

Majorana Fermion Mean-Field Theories of Kitaev Quantum Spin Liquids

We determine the phase diagrams of anisotropic Kitaev-Heisenberg models on the honeycomb lattice using parton mean-field theories based on different Majorana fermion representations of the $S=1/2$ spin operators. Firstly, we use a two-dimensional Jordan-Wigner transformation (JWT) involving a semi-infinite snake string operator. In order to ensure that the fermionized Hamiltonian remains local we consider the limit of extreme Ising exchange anisotropy in the Heisenberg sector. Secondly, we use the conventional Kitaev representation in terms of four Majorana fermions subject to local constraints, which we enforce through Lagrange multipliers. For both representations we self-consistently decouple the interaction terms in the bond and magnetization channels and determine the phase diagrams as a function of the anisotropy of the Kitaev couplings and the relative strength of the Ising exchange. While both mean-field theories produce identical phase boundaries for the topological phase transition between the gapless and gapped Kitaev quantum spin liquids, the JWT fails to correctly describe the the magnetic instability and finite-temperature behavior. Our results show that the magnetic phase transition is first order at low temperatures but becomes continuous above a certain temperature. At this energy scale we also observe a finite temperature crossover on the quantum spin-liquid side, from a fractionalized paramagnet at low temperatures, in which gapped flux excitations are frozen out, to a conventional paramagnet at high temperatures.

cond-mat.str-el↗
arXiv2022

Stability of the N\'eel quantum critical point in the presence of Dirac fermions

We investigate the stability of the N\'eel quantum critical point of two-dimensional quantum antiferromagnets, described by a non-linear $\sigma$ model (NL$\sigma$M), in the presence of a Kondo coupling to $N_f$ flavours of two-component Dirac fermion fields. The long-wavelength order parameter fluctuations are subject to Landau damping by electronic particle-hole fluctuations. Using momentum-shell RG, we demonstrate that the Landau damping is weakly irrelevant at the N\'eel quantum critical point, despite the fact that the corresponding self-energy correction dominates over the quadratic gradient terms in the IR limit. In the ordered phase, the Landau damping increases under the RG, indicative of damped spin-wave excitations. Although the Kondo coupling is weakly relevant, sufficiently strong Landau damping renders the N\'eel quantum critical point quasi-stable for $N_f\ge 4$ and thermodynamically stable for $N_f<4$. In the latter case, we identify a new multi-critical point which describes the transition between the N\'eel critical and Kondo run-away regimes. The symmetry breaking at this fixed point results in the opening of a gap in the Dirac fermion spectrum. Approaching the multi-critical point from the disordered phase, the fermionic quasiparticle residue vanishes, giving rise to non-Fermi-liquid behavior.

cond-mat.str-el↗
arXiv2021

A new look at the entanglement entropy of a single interval in a 2d CFT

In this note, I revisit the problem of computing the entanglement entropy of a single interval in the ground state of a 2d CFT. I write the leading-order result in three different ways: once by doing the replica trick with the $n$-replicated cylinder partition function computed in the ``closed string channel"; once by computing the $n$-replicated cylinder partition function in the ``open string channel", where the entanglement entropy can be related to the density of states in a boundary CFT defined on the interval; and for holographic CFTs, once as the log of a Plancherel measure for a certain noncompact quantum group. I comment on the implications for what the Ryu-Takayanagi area term might be counting in holographic CFTs.

hep-th↗
arXiv2021

Entanglement entropy in Jackiw-Teitelboim gravity with matter

In this paper, I study the entanglement entropy in Hartle-Hawking states of JT gravity set up by a Euclidean path integral with an operator inserted somewhere along the Euclidean boundary. I show that the entanglement entropy between the dual SYK models can be written to leading order in $1/G_N$ in a way that resembles a formula for entanglement entropy in a compact lattice gauge theory, but where the link that we appear to compute the entanglement across depends on where we put the operator along the Euclidean boundary. I then describe a toy model where we can write down an explicit family of density matrices whose von Neumann entropies exhibit such a phase transition as we tune an interpolating parameter. This behavior relies on a certain superselection sector structure and I comment on the possibility that the gravitational result comes from a similar emergent superselection sector structure in SYK/JT holography. The main implication is that each copy of the SYK model should then contain Hilbert space sectors labeled by multiple $SL(2,R)$ representations whose entanglement gives rise to space.

hep-th↗
arXiv2018

Comments on Defining Entanglement Entropy

We revisit the issue of defining the entropy of a spatial region in a broad class of quantum theories. In theories with explicit regularizations, working within an elementary but general algebraic framework applicable to matter and gauge theories alike, we give precise path integral expressions for three known types of entanglement entropy that we call full, distillable, and gauge-invariant. For a class of gauge theories that do not necessarily have a regularization in our framework, including Chern-Simons theory, we describe a related approach to defining entropies based on locally extending the Hilbert space at the entangling edge, and we discuss its connections to other calculational prescriptions. Based on results from both approaches, we conjecture that it is always the full entanglement entropy that is calculated by standard holographic techniques in strongly coupled conformal theories.

hep-th↗
arXiv2018

Entanglement entropy in Jackiw-Teitelboim Gravity

I show that the black hole entropy associated to an $AdS_2$ wormhole is an entanglement edge term related to a natural measure on the gauge group in the $SL(2)$ gauge theory formulation of $1+1d$ Jackiw-Teitelboim gravity. I comment on what the entropy appears to be counting.

hep-th↗
arXiv2017

Ryu-Takayanagi Area as an Entanglement Edge Term

By comparing entanglement in emergent gauge theories to the bulk in AdS/CFT, I suggest that the Ryu-Takayanagi area term is an entanglement edge term related to a natural measure on the gauge group. The main technical result in this paper is an argument why the "extended Hilbert space" definition of entanglement entropy in a lattice gauge theory is applicable to an emergent gauge theory.

hep-th↗
arXiv2016

A Toy Model of Entwinement

Entwinement is the entanglement entropy of a subset of gauge-variant degrees of freedom in a certain twisted state of an orbifold CFT, defined by embedding the state in a larger Hilbert space with some gauge constraints removed. We propose an intrinsically gauge-invariant, algebraic definition of entwinement. Our main piece of evidence is a computation showing that, in a spin system that resembles the orbifold CFT, the analog of entwinement is the entanglement entropy of a gauge-invariant subalgebra, which we identify. We review why entwinement is relevant for the conjecture that entanglement builds spacetime.

hep-th↗
arXiv2016

Gravitational Positive Energy Theorems from Information Inequalities

In this paper we argue that classical, asymptotically AdS spacetimes that arise as states in consistent ultraviolet completions of Einstein gravity coupled to matter must satisfy an infinite family of positive energy conditions. To each ball-shaped spatial region $B$ of the boundary spacetime, we can associate a bulk spatial region $\Sigma_B$ between $B$ and the bulk extremal surface $\tilde{B}$ with the same boundary as $B$. We show that there exists a natural notion of a gravitational energy for every such region that is non-negative, and non-increasing as one makes the region smaller. The results follow from identifying this gravitational energy with a quantum relative entropy in the associated dual CFT state. The positivity and monotonicity properties of the gravitational energy are implied by the positivity and monotonicity of relative entropy, which holds universally in all quantum systems.

hep-th↗
arXiv2015

Bulk Locality from Entanglement in Gauge/Gravity Duality

Gauge/gravity duality posits an equivalence between certain strongly coupled quantum field theories and theories of gravity with negative cosmological constant in a higher number of spacetime dimensions. The map between the degrees of freedom on the two sides is non-local and incompletely understood. I describe recent work towards characterizing this map using entanglement in the QFT, where near the dual AdS boundary, the classical energy density at a point in the bulk is stored in the relative entropies of boundary subregions whose homologous minimal surfaces pass through the bulk point. I also derive bulk classical energy conditions near the AdS boundary from entanglement inequalities in the CFT. This is based on the paper [1] with Matilde Marcolli, Hirosi Ooguri and Bogdan Stoica. More generally, in recent years, there has appeared some evidence that quantum entanglement is responsible for the emergence of spacetime. I review and comment on the state of these developments.

hep-th↗
arXiv2014

Tomography from Entanglement

The Ryu-Takayanagi formula relates the entanglement entropy in a conformal field theory to the area of a minimal surface in its holographic dual. We show that this relation can be inverted for any state in the conformal field theory to compute the bulk stress-energy tensor near the boundary of the bulk spacetime, reconstructing the local data in the bulk from the entanglement on the boundary. We also show that positivity, monotonicity, and convexity of the relative entropy for small spherical domains between the reduced density matrices of any state and of the ground state of the conformal field theory, follow from positivity conditions on the bulk matter energy density. We discuss an information theoretical interpretation of the convexity in terms of the Fisher metric.

hep-th↗
arXiv2014

(0,2) ADE Models From Four Dimensions

Four dimensional N=1 supersymmetric gauge theories with unitary gauge groups and matter in the adjoint and fundamental representations give rise to a series of non-trivial fixed points with an ADE classification. Many of these models exhibit generalizations of Seiberg duality. Upon compactification on a two-torus with suitable background fields for global U(1) symmetries, they flow at long distances to two dimensional theories with (0,2) supersymmetry. We study these theories, focusing on the two dimensional analogs of the four dimensional dualities.

hep-th↗
arXiv2014

Exceptional N=1 Duality

Four dimensional N=1 supersymmetric gauge theory with gauge group SU(N_c) and matter in the adjoint and fundamental representations gives rise to a series of fixed points with an ADE classification. The A and D series exhibit generalizations of Seiberg duality. We propose a similar duality for the E7 theory.

hep-th↗