pith. sign in

arxiv: 2606.24671 · v1 · pith:P24WXXXHnew · submitted 2026-06-23 · ❄️ cond-mat.dis-nn

Anderson localization on the Bethe lattice

Pith reviewed 2026-06-25 22:09 UTC · model grok-4.3

classification ❄️ cond-mat.dis-nn
keywords Anderson localizationBethe latticeresolvent formalismcavity methodlocal density of statesdisorder-driven transitiondirected polymers
0
0 comments X

The pith

The full probability distribution of the local density of states serves as the order parameter for the Anderson localization transition on the Bethe lattice.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The notes review how Anderson localization on the Bethe lattice becomes analytically tractable through the resolvent formalism and cavity method. They establish that the entire probability distribution of the local density of states, rather than its average value, distinguishes the localized and delocalized phases. This approach yields the critical disorder strength, the localization length on the insulating side, and the correlation volume on the delocalized side. A reader cares because the Bethe lattice supplies a solvable infinite-dimensional limit relevant to many-body localization problems.

Core claim

On the Bethe lattice the Anderson transition is captured by the statistical properties of Green's functions. The cavity self-consistent equations for the diagonal resolvent are derived and solved both numerically by population dynamics and analytically in the large-connectivity limit; the resulting fixed-point distribution of the local density of states functions as the natural order parameter, giving access to critical disorder, localization length, and correlation volume while also linking the problem to directed polymers in random media.

What carries the argument

The cavity self-consistent equations for the diagonal resolvent, whose fixed-point solution is the probability distribution of the local density of states.

If this is right

  • The critical disorder strength follows from the point at which the distribution of local density of states changes character.
  • The localization length in the insulating phase is obtained from the typical decay of the resolvent.
  • The correlation volume that controls the delocalized phase is extracted from the same distribution near criticality.
  • Rare resonant paths and associated freezing map the localization problem onto directed polymers in random media.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same distribution-based order parameter could be monitored in numerical studies of finite-dimensional lattices or many-body systems to locate transitions without relying on averaged quantities.
  • Connections to directed polymers suggest that rare-region effects might dominate near the transition in real materials, offering a route to predict Griffiths-like singularities.
  • Analytical large-connectivity results could serve as benchmarks for approximate theories of localization on graphs with different topologies.

Load-bearing premise

The Bethe lattice realizes the infinite-dimensional limit of the Anderson problem and serves as a useful mean-field description of localization in many-body configuration spaces.

What would settle it

A direct numerical computation on large finite Bethe lattices showing that the full distribution of the local density of states fails to separate localized from delocalized regimes would falsify the central claim.

Figures

Figures reproduced from arXiv: 2606.24671 by Angelica Vecchiarelli, Dafne Prado Bandeira, Marco Tarzia.

Figure 1
Figure 1. Figure 1: Schematic representation of the configuration space of the spin chain (Eq. (21)) for L = 3 in the basis of the simultaneous eigenstates of the {σ z i } operators. Each vertex corresponds to one of the 2 3 = 8 spin configurations. To each vertex, a random energy is assigned (indicated by color) containing linear combinations of the random fields hi , weighted by the spin values of that specific configuratio… view at source ↗
Figure 2
Figure 2. Figure 2: (Left) For a finite system, G(z) is analytic except at a discrete set of points on the real axis corresponding to the eigenvalues λn, where it exhibits simple poles. (Right) In the thermodynamic limit (N → ∞), these discrete poles may merge into a continuous branch cut along the real axis. The physical Green’s functions are recovered by approaching the real axis from the complex plane through the regulariz… view at source ↗
Figure 3
Figure 3. Figure 3: Energy spectrum of the system in the extended phase. Each horizontal line represents an eigenstate at energy λn, with its oscillatory pattern indicating the spatial structure. The red marker shows that the amplitude of the wavefunction is uniformly of order O(1/N). may remain pure point (localized phase) or develop an absolutely continuous component (delocalized phase) in the thermodynamic limit. However, … view at source ↗
Figure 4
Figure 4. Figure 4: Integrated local density of states Ri(E) in the extended phase. Both the step height (proportional to |ψn(i)| 2 ) and the level spacing are of order O(1/N). In the thermodynamic limit, these discrete steps merge into a smooth, continuous function. In the thermodynamic limit N → ∞, the discrete steps in Ri(E) become infinitely fine, and the function converges to a smooth, continuously increasing function of… view at source ↗
Figure 5
Figure 5. Figure 5: Energy spectrum in the localized regime. Each eigenstate is confined to a region around its localization center. At site i, the local density of states receives significant contributions only from eigenstates whose centers lie within ∼ ξ of i; distant states contribute negligibly. 3.2.2. Localized eigenstates The localized phase presents a notably different picture. Here, eigenstates are exponentially conf… view at source ↗
Figure 6
Figure 6. Figure 6: Integrated local density of states Ri(E) in the localized phase. Most steps are exponentially small (|ψn(i)| 2 ∼ e −L/ξ), corresponding to eigenstates localized far (at a distance L) from site i. Rare large jumps of order one occur when an eigenstate is centered near i. This irregular structure persists even in the thermodynamic limit. and is associated with the emergence of a branch cut in the resolvent. … view at source ↗
Figure 7
Figure 7. Figure 7: Cayley tree of degree k = 3 up to the third generation. Locally, it is identical to the neighborhood of the RRG and of the BL of same degree. Each concentric circle shows a successive generation of nodes, where the distance ℓ from the central root increases by one at each layer. properties. In contrast, the Bethe Lattice is an infinite tree which has no boundaries and is translationally invariant, i.e., al… view at source ↗
Figure 8
Figure 8. Figure 8: Schematic representation of the cavity method factorization. Thus, the matrix elements of the resolvent can be expressed as correlation functions of the Gaussian fields ϕi with the action S[ϕ] = − i 2 X N i=1 (z − εi)ϕ 2 i + i X ⟨i,j⟩ t ϕiϕj , (73) where i denotes the sum over all nodes and ⟨i, j⟩ the sum over nearest-neighbor pairs on the lattice. The elements of the resolvent can thus be obtained as −iGa… view at source ↗
Figure 9
Figure 9. Figure 9: Branching structure entering the computation of the off-diagonal resolvent G0r(z) on the Bethe lattice. The final section of the unique path between nodes 0 and r is shown horizontally. The final node, r, has k neighbors to be integrated out that are not part of the path connecting it to node 0, while all the intermediary nodes in the path, as r − 1, have k − 1 such neighbors. starting point is the fully l… view at source ↗
Figure 10
Figure 10. Figure 10: Probability distribution of Im G in the localized phase for different values of the regulator η. The distributions are obtained using population dynamics (see Sec.7) for a population size Ω = 225, and fixed energy E = 0.0 and disorder strength W = 30.0, on a Bethe lattice with branching k = 2. As η → 0 +, the probability distribution becomes singular: the lower cutoff moves to zero, while the upper cutoff… view at source ↗
Figure 11
Figure 11. Figure 11: Probability distribution of Im G in the delocalized phase for different values of the disorder strength W. The distributions are obtained using population dynamics (see Sec.7) for a population size Ω = 225, and fixed energy E = 0.0 and no regulator η = 0.0, on a Bethe lattice with branching k = 2, for which the critical disorder is Wc ≈ 18. referred to as the correlation volume. To clarify why this identi… view at source ↗
Figure 12
Figure 12. Figure 12: Schematic sketch of the spatial profile of the amplitude of typical delocalized eigenstates close to localization threshold, resulting of the hybridization of N/Λc resonant peaks, each one occupying a volume Λc. approaching the transition [38, 42, 31], see Eq. (127) below. As a consequence, all moments of Im G remain finite. In particular, high-order moments (q > 1/2) are dominated by the upper cutoff and… view at source ↗
Figure 13
Figure 13. Figure 13: Schematic phase diagram of the Anderson model for the Bethe Lattice for k = 2 in the W-E plane [132, 133, 134]. The inner curve marks the mobility edge separating localized and extended states. The straight lines indicate the edges of the support of the density of states, E = ±(2√ k+W/2); the shaded outer regions therefore correspond to energies outside the spectrum of the model. The vertical dashed line … view at source ↗
Figure 14
Figure 14. Figure 14: Illustration of the largest eigenvalue of the integral operator (Eq. (106)) as a function of β ∈ [0, 1]. The green curve corresponds to W < Wc (unstable localized phase). The red curve represents the critical point W = Wc, and the other curves correspond to W > Wc (stable localized phase). The intersections λβ = 1 determine the value of β for each disorder strength. the largest eigenvalue is symmetric aro… view at source ↗
Figure 15
Figure 15. Figure 15: Sketch of the critical behavior near the Anderson transition on the Bethe lattice [27, 6, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43]. On the localized side, W ≳ Wc, the length ξ diverges as ξ ∼ (W − Wc) −1 (Eq. (124)), the tail exponent approaches its critical value as β − 1/2 ∼ √ W − Wc, Eq. (125), and the IPR approaches its finite critical value as I2(W) − I (c) 2 ∼ √ W − Wc (see Sec. 7… view at source ↗
Figure 16
Figure 16. Figure 16: Distribution P(ImG) of the imaginary part of the cavity Green’s function for several values W ∈ [13, 17.3]. The lines correspond to fits using Eq. (155) to extract the correlation volume (see text). This large-deviation approach was applied in Ref. [139] to compute the distribution of the cavity Green’s function for the Anderson model on the Bethe lattice with [PITH_FULL_IMAGE:figures/full_fig_p055_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Left: Probability distribution Q(0, mˆ )/ ⟨ρ⟩ obtained by conditioning on m = 0. Middle: Exponent β of the power-law tails as a function of W. The dashed curve shows the prediction of Eq. (125) near Wc from Ref. [31]. Right: Generalized inverse participation ratios Ip (for η = 0 and N → ∞) as a function of W in the localized phase for p = 1.4, p = 2, and p = 4. The dashed lines are fits of the form Ip ≃ I… view at source ↗
read the original abstract

Anderson localization is a paradigmatic disorder-driven quantum phase transition in which interference effects can completely suppress wave propagation. Its formulation on the Bethe lattice -- an infinite tree with fixed coordination number -- provides a unique setting in which the transition becomes analytically tractable. Owing to its exponential volume growth, the Bethe lattice realizes the infinite-dimensional limit of the Anderson problem and also serves as a useful mean-field-like description of localization in many-body configuration spaces. These lecture notes provide a pedagogical review of Anderson localization on the Bethe lattice, focusing on the resolvent (Green's function) formalism and the cavity approach. We discuss the main diagnostics of localization and show how they are encoded in the statistical properties of Green's functions, with the full probability distribution of the local density of states emerging as the natural order parameter of the transition. We derive the cavity self-consistent equations for the diagonal resolvent and review both their numerical solution by population dynamics and their analytical treatment in the large-connectivity limit. This framework gives access to the critical disorder and the critical behavior near the transition, including the localization length in the insulating phase and the correlation volume that controls the delocalized phase. We also discuss the connection with directed polymers in random media and its interpretation in terms of rare resonant paths and freezing phenomena. These notes originated from a series of lectures delivered by one of the authors at the Oropa Summer School on "Fundamental Problems in Statistical Physics XVI" (July 2025).

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. This manuscript is a set of lecture notes that review Anderson localization on the Bethe lattice. It employs the resolvent formalism and cavity method to derive self-consistent equations whose solution is the full probability distribution of the local density of states (the order parameter). The notes cover numerical solution via population dynamics, analytic treatment in the large-connectivity limit, the critical disorder strength, the localization length on the insulating side, the correlation volume on the delocalized side, and the mapping to directed polymers in random media that interprets rare resonant paths and freezing.

Significance. If the derivations and numerical procedures are presented accurately, the notes would constitute a compact pedagogical resource for a technique that remains central to mean-field treatments of localization, both for the infinite-dimensional Anderson model and as an approximation for many-body configuration spaces. The explicit identification of the full distribution of the local density of states as the order parameter, together with the discussion of population dynamics and the large-connectivity limit, reproduces standard but non-trivial results from the literature in a single, accessible source.

minor comments (3)
  1. [Abstract] Abstract, final paragraph: the origin of the notes (Oropa Summer School, July 2025) is stated; a short sentence indicating which parts of the text are original derivations versus direct reproductions of earlier literature would help readers assess novelty.
  2. The manuscript repeatedly refers to 'the cavity self-consistent equations' without an early numbered equation that collects the full set of distributional equations; adding such an equation (e.g., after the first derivation) would improve readability.
  3. Section on the large-connectivity limit: the scaling of the critical disorder with coordination number is discussed, but the precise asymptotic expression is not written as a numbered equation; doing so would make the analytic result easier to cite.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of our lecture notes and for recommending acceptance. The report contains no major comments requiring a point-by-point reply.

Circularity Check

0 steps flagged

No significant circularity: pedagogical review of established results

full rationale

The paper is explicitly a review of prior derivations in the cavity-method literature for Anderson localization on the Bethe lattice. The central claim that the full distribution of the local density of states is the natural order parameter follows directly from the structure of the self-consistent resolvent equations (which close on the distribution rather than moments), a standard feature already present in the cited foundational works. No new parameters are fitted within the paper, no self-citation chain is load-bearing for a novel result, and no ansatz or uniqueness theorem is smuggled in. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

As a review paper, the work relies on standard assumptions from the prior literature on Anderson localization and cavity methods without introducing new free parameters, axioms, or invented entities.

pith-pipeline@v0.9.1-grok · 5795 in / 1006 out tokens · 12694 ms · 2026-06-25T22:09:36.154010+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

159 extracted references · 66 canonical work pages

  1. [1]

    5 for the Anderson model on the Bethe lattice must be interpreted as self-consistent integral equations for the probability CONTENTS52 distribution of the cavity Green’s functions

    Population Dynamics algorithm(s) As discussed above, the cavity equations derived in Sec. 5 for the Anderson model on the Bethe lattice must be interpreted as self-consistent integral equations for the probability CONTENTS52 distribution of the cavity Green’s functions. These equations are given by Eqs. (87) and (88). The local density of states, the inve...

  2. [2]

    Connection with Derrida’s directed polymers in random media Before concluding this review, we would like to discuss the close relationship between Anderson localization on the Bethe lattice and a paradigmatic problem of classical disordered systems, namely the freezing glass transition of directed polymers in random media (DPRM), first introduced and stud...

  3. [3]

    Plugging the expression ofm ⋆ into the replicated free energy, one recovers the expression (167). It is easy to check that for the particular case in whichγ(ω)is a Gaussian distribution with zero mean and varianceσ2 ω one recovers the results (165) and (166)— withβ ⋆ = √ 2 lnk/σ ω—corresponding to the approximation in which one neglects the correlations b...

  4. [4]

    Fundamental Problems in Statistical Physics XVI

    Conclusions and perspectives 9.1. Summary More than six decades after Philip W. Anderson’s seminal discovery [1], the problem of quantum localization driven by disorder continues to reveal new frontiers, challenges, and subtleties. This pedagogical review has been dedicated to exploring this phenomenon within a unique and highly singular framework: the in...

  5. [5]

    Rev.109(5) 1492–1505 URLhttps://link.aps.org/doi/10.1103/ PhysRev.109.1492

    Anderson P W 1958Phys. Rev.109(5) 1492–1505 URLhttps://link.aps.org/doi/10.1103/ PhysRev.109.1492

  6. [6]

    Bloch F 1928Zeitschrift für Physik52555–600

  7. [7]

    Drude P 1900Annalen der Physik306566–613 (Preprinthttps://onlinelibrary.wiley.com/ doi/pdf/10.1002/andp.19003060312) URLhttps://onlinelibrary.wiley.com/doi/abs/ 10.1002/andp.19003060312

  8. [8]

    Sommerfeld A 1927Naturwissenschaften15825–832

  9. [9]

    Lee P A and Ramakrishnan T V 1985Rev. Mod. Phys.57(2) 287–337 URLhttps://link.aps. org/doi/10.1103/RevModPhys.57.287

  10. [10]

    Evers F and Mirlin A D 2008Rev. Mod. Phys.80(4) 1355–1417 URLhttps://link.aps.org/ doi/10.1103/RevModPhys.80.1355 CONTENTS69

  11. [11]

    Lagendijk A, Tiggelen B v and Wiersma D S 2009Physics today6224–29 URLhttps: //physicstoday.aip.org/features/fifty-years-of-anderson-localization

  12. [12]

    Scardicchio A and Thiery T 2017 Perturbation theory approaches to anderson and many-body localization: some lecture notes (Preprint1710.01234) URLhttps://arxiv.org/abs/1710. 01234

  13. [13]

    Pascazio S, Scardicchio A and Tarzia M 2023 Anderson localization on the bethe latticeSpin Glass Theory and Far Beyond: Replica Symmetry Breaking After 40 Years(World Scientific) pp 335–352

  14. [14]

    Mott N 1967Advances in Physics1649–144 (Preprinthttps://doi.org/10.1080/ 00018736700101265) URLhttps://doi.org/10.1080/00018736700101265

  15. [15]

    Abrahams E, Anderson P W, Licciardello D C and Ramakrishnan T V 1979Physical Review Letters42673

  16. [16]

    Slevin K and Ohtsuki T 1999Physical Review Letters82382–385 ISSN 1079-7114 URL http://dx.doi.org/10.1103/PhysRevLett.82.382

  17. [17]

    Slevin K and Ohtsuki T 2014New Journal of Physics16015012 ISSN 1367-2630 URLhttp: //dx.doi.org/10.1088/1367-2630/16/1/015012

  18. [18]

    Tarquini E, Biroli G and Tarzia M 2017Physical Review B95094204

  19. [19]

    Hikami S 1992Progress of Theoretical Physics Supplement107213–227

  20. [20]

    Wegner F 1979Zeitschrift für Physik B Condensed Matter35207–210

  21. [21]

    Schäfer L and Wegner F 1980Zeitschrift für Physik B Condensed Matter38113–126

  22. [22]

    John S 1984Phys. Rev. Lett.53(22) 2169–2172 URLhttps://link.aps.org/doi/10.1103/ PhysRevLett.53.2169

  23. [23]

    1080/13642818508240619) URLhttps://doi.org/10.1080/13642818508240619

    Anderson P W 1985Philosophical Magazine B52505–509 (Preprinthttps://doi.org/10. 1080/13642818508240619) URLhttps://doi.org/10.1080/13642818508240619

  24. [24]

    Billy J, Josse V, Zuo Z, Bernard A, Hambrecht B, Lugan P, Clément D, Sanchez-Palencia L, Bouyer P and Aspect A 2008Nature453891–894

  25. [25]

    Roati G, D’Errico C, Fallani L, Fattori M, Fort C, Zaccanti M, Modugno G, Modugno M and Inguscio M 2008Nature453895–898

  26. [26]

    Kondov S S, McGehee W R, Zirbel J J and DeMarco B 2011Science33466–68 ISSN 1095-9203 URLhttp://dx.doi.org/10.1126/science.1209019

  27. [27]

    Jendrzejewski F, Bernard A, Müller K, Cheinet P, Josse V, Piraud M, Pezzé L, Sanchez- Palencia L, Aspect A and Bouyer P 2012Nature Physics8398–403 ISSN 1745-2481 URL http://dx.doi.org/10.1038/nphys2256

  28. [28]

    Semeghini G, Landini M, Castilho P, Roy S, Spagnolli G, Trenkwalder A, Fattori M, Inguscio M and Modugno G 2015Nature Physics11554–559 ISSN 1745-2481 URLhttp://dx.doi.org/ 10.1038/nphys3339

  29. [29]

    Chabé J, Lemarié G, Grémaud B, Delande D, Szriftgiser P and Garreau J C 2008Physical Review Letters101ISSN 1079-7114 URLhttp://dx.doi.org/10.1103/PhysRevLett.101.255702

  30. [30]

    Hu H, Strybulevych A, Page J H, Skipetrov S E and van Tiggelen B A 2008Nature Physics4 945–948 ISSN 1745-2481 URLhttp://dx.doi.org/10.1038/nphys1101

  31. [31]

    Abou-Chacra R, Thouless D J and Anderson P W 1973Journal of Physics C: Solid State Physics 61734 URLhttps://doi.org/10.1088/0022-3719/6/10/009

  32. [33]

    Rizzo T and Tarzia M 2024Phys. Rev. B110(18) 184210 URLhttps://link.aps.org/doi/ 10.1103/PhysRevB.110.184210

  33. [34]

    Tikhonov K S and Mirlin A D 2019Phys. Rev. B99(2) 024202 URLhttps://link.aps.org/ doi/10.1103/PhysRevB.99.024202

  34. [35]

    Tikhonov K S and Mirlin A D 2019Phys. Rev. B99(21) 214202 URLhttps://link.aps.org/ doi/10.1103/PhysRevB.99.214202

  35. [36]

    Efetov K 1985Sov. Phys. JETP61606–617 URLhttps://jetp.ras.ru/cgi-bin/dn/e_061_ CONTENTS70 03_0606.pdf

  36. [37]

    Efetov K 1987Sov. Phys. JETP65360–370 URLhttps://jetp.ras.ru/cgi-bin/dn/e_065_ 02_0360.pdf

  37. [38]

    Efetov K 1987Sov. Phys. JETP66634–642 URLhttps://jetp.ras.ru/cgi-bin/dn/e_066_ 03_0634.pdf

  38. [39]

    Zirnbauer M R 1986Phys. Rev. B34(9) 6394–6408 URLhttps://link.aps.org/doi/10.1103/ PhysRevB.34.6394

  39. [42]

    sciencedirect.com/science/article/pii/055032139190028V

    Mirlin A D and Fyodorov Y V 1991Nuclear Physics B366507–532 URLhttps://www. sciencedirect.com/science/article/pii/055032139190028V

  40. [43]

    Mirlin A and Fyodorov Y V 1991Journal of Physics A: Mathematical and General242273 URL https://doi.org/10.1088/0305-4470/24/10/016

  41. [44]

    Fyodorov Y V and Mirlin A D 1991Phys. Rev. Lett.67(15) 2049–2052 URLhttps://link.aps. org/doi/10.1103/PhysRevLett.67.2049

  42. [45]

    Fyodorov Y V, Mirlin A D and Sommers H J 1992J. Phys. I France21571–1605 ISSN 1286-4862 URLhttp://dx.doi.org/10.1051/jp1:1992229

  43. [46]

    Mirlin A D and Fyodorov Y V 1994Journal de Physique I4655–673

  44. [47]

    Mirlin A D and Fyodorov Y V 1994Phys. Rev. Lett.72(4) 526–529 URLhttps://link.aps. org/doi/10.1103/PhysRevLett.72.526

  45. [48]

    Gornyi I V, Mirlin A D and Polyakov D G 2005Phys. Rev. Lett.95(20) 206603 URLhttps: //link.aps.org/doi/10.1103/PhysRevLett.95.206603

  46. [49]

    Basko D, Aleiner I and Altshuler B 2006Ann. Phys. (N. Y.)3211126–1205 ISSN 0003-4916 URLhttp://dx.doi.org/10.1016/j.aop.2005.11.014

  47. [50]

    Nandkishore R and Huse D A 2015Annu. Rev. Condens. Mat. Phys.615–38 ISSN 1947-5462 URLhttps://doi.org/10.1146/annurev-conmatphys-031214-014726

  48. [51]

    Abanin D A, Altman E, Bloch I and Serbyn M 2019Rev. Mod. Phys.91(2) 021001 URL https://link.aps.org/doi/10.1103/RevModPhys.91.021001

  49. [52]

    Physique19498–525 ISSN 1878-1535 URL http://dx.doi.org/10.1016/j.crhy.2018.03.003

    Alet F and Laflorencie N 2018Comptes Rendus. Physique19498–525 ISSN 1878-1535 URL http://dx.doi.org/10.1016/j.crhy.2018.03.003

  50. [53]

    Sierant P, Lewenstein M, Scardicchio A, Vidmar L and Zakrzewski J 2025Reports on Progress in Physics88026502 ISSN 1361-6633 URLhttp://dx.doi.org/10.1088/1361-6633/ad9756

  51. [54]

    Huse D A, Nandkishore R and Oganesyan V 2014Phys. Rev. B90ISSN 1550-235X URL http://dx.doi.org/10.1103/PhysRevB.90.174202

  52. [55]

    Altshuler B L, Gefen Y, Kamenev A and Levitov L S 1997Physical Review Letters782803–2806 ISSN 1079-7114 URLhttp://dx.doi.org/10.1103/PhysRevLett.78.2803

  53. [56]

    Tikhonov K and Mirlin A 2021Annals of Physics435168525 ISSN 0003-4916 URLhttp: //dx.doi.org/10.1016/j.aop.2021.168525

  54. [57]

    Tikhonov K S and Mirlin A D 2021Phys. Rev. B103(6) 064204 URLhttps://link.aps.org/ doi/10.1103/PhysRevB.103.064204

  55. [58]

    Tarzia M 2020Physical Review B102ISSN 2469-9969 URLhttp://dx.doi.org/10.1103/ PhysRevB.102.014208

  56. [59]

    De Luca A and Scardicchio A 2013EPL (Europhysics Letters)10137003 ISSN 1286-4854 URL http://dx.doi.org/10.1209/0295-5075/101/37003

  57. [60]

    Biroli G and Tarzia M 2017Phys. Rev. B96(20) 201114(R) URLhttps://link.aps.org/doi/ 10.1103/PhysRevB.96.201114

  58. [61]

    Biroli G and Tarzia M 2020Phys. Rev. B102(6) 064211 URLhttps://link.aps.org/doi/10. 1103/PhysRevB.102.064211

  59. [62]

    Logan D E and Welsh S 2019Physical Review B99ISSN 2469-9969 URLhttp://dx.doi.org/ CONTENTS71 10.1103/PhysRevB.99.045131

  60. [63]

    García-Mata I, Martin J, Giraud O, Georgeot B, Dubertrand R and Lemarié G 2022Phys. Rev. B106(21) 214202 URLhttps://link.aps.org/doi/10.1103/PhysRevB.106.214202

  61. [64]

    Herre J N, Karcher J F, Tikhonov K S and Mirlin A D 2023Physical Review B108ISSN 2469-9969 URLhttp://dx.doi.org/10.1103/PhysRevB.108.014203

  62. [65]

    org/abs/1211.7334

    Biroli G, Ribeiro-Teixeira A C and Tarzia M 2012 Difference between level statistics, ergodicity and localization transitions on the bethe lattice (Preprint1211.7334) URLhttps://arxiv. org/abs/1211.7334

  63. [66]

    Kravtsov V, Altshuler B and Ioffe L 2018Annals of Physics389148–191 ISSN 0003-4916 URL http://dx.doi.org/10.1016/j.aop.2017.12.009

  64. [67]

    Pino M 2020Phys. Rev. Res.2(4) 042031 URLhttps://link.aps.org/doi/10.1103/ PhysRevResearch.2.042031

  65. [68]

    Bera S, De Tomasi G, Khaymovich I M and Scardicchio A 2018Phys. Rev. B98(13) 134205 URLhttps://link.aps.org/doi/10.1103/PhysRevB.98.134205

  66. [69]

    De Tomasi G, Bera S, Scardicchio A and Khaymovich I M 2020Phys. Rev. B101(10) 100201 URLhttps://link.aps.org/doi/10.1103/PhysRevB.101.100201

  67. [70]

    Wormald N Cet al.1999London mathematical society lecture note series239–298

  68. [71]

    Tikhonov K S, Mirlin A D and Skvortsov M A 2016Phys. Rev. B94(22) 220203 URL https://link.aps.org/doi/10.1103/PhysRevB.94.220203

  69. [72]

    García-Mata I, Giraud O, Georgeot B, Martin J, Dubertrand R and Lemarié G 2017Phys. Rev. Lett.118(16) 166801 URLhttps://link.aps.org/doi/10.1103/PhysRevLett.118.166801

  70. [73]

    Tarquini E, Biroli G and Tarzia M 2016Phys. Rev. Lett.116(1) 010601 URLhttps://link. aps.org/doi/10.1103/PhysRevLett.116.010601

  71. [74]

    Biroli G and Tarzia M 2018 Delocalization and ergodicity of the anderson model on bethe lattices (Preprint1810.07545) URLhttps://arxiv.org/abs/1810.07545

  72. [75]

    Baroni M, Lorenzana G G, Rizzo T and Tarzia M 2024Phys. Rev. B109(17) 174216 URL https://link.aps.org/doi/10.1103/PhysRevB.109.174216

  73. [77]

    Efetov K 1983Advances in Physics3253–127 (Preprinthttps://doi.org/10.1080/ 00018738300101531) URLhttps://doi.org/10.1080/00018738300101531

  74. [78]

    Rep.326259–382 ISSN 0370-1573 URLhttps://www.sciencedirect

    Mirlin A D 2000Phys. Rep.326259–382 ISSN 0370-1573 URLhttps://www.sciencedirect. com/science/article/pii/S0370157399000915

  75. [79]

    Sierant P, Lewenstein M and Scardicchio A 2023SciPost Physics15045

  76. [80]

    García-Mata I, Martin J, Dubertrand R, Giraud O, Georgeot B and Lemarié G 2020Phys. Rev. Res.2(1) 012020(R) URLhttps://link.aps.org/doi/10.1103/PhysRevResearch.2. 012020

  77. [81]

    Vanoni C, Altshuler B L, Kravtsov V E and Scardicchio A 2024Proceedings of the National Academy of Sciences121e2401955121

  78. [82]

    Sonner M, Tikhonov K and Mirlin A 2017Physical Review B96214204

  79. [83]

    Tikhonov K S and Mirlin A D 2016Physical Review B94184203

  80. [84]

    Derrida B and Spohn H 1988Journal of Statistical Physics51817–840

Showing first 80 references.