Pith. sign in

REVIEW 4 major objections 5 minor 40 references

The paper argues that requiring well-defined QCD-induced mass mixing forces the ALP masses into a hierarchical ladder and splits the decay constants into two populations separated by an empty window, making the axiverse's ordering a low-ene

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

The paper asserts that the requirement of well-defined QCD-induced mixing alone forces ALP masses into a hierarchical spacing and decay constants into two separated populations, but it does not demonstrate the underlying derivations.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection The paper's two claimed emergent patterns are both baked into the setup—the decay constant gap is an input of Eq. (5), and the 1/2 mass ratio is asserted, not derived. the 4 major comments →

arxiv 2608.03678 v1 pith:STFXTXC7 submitted 2026-08-04 hep-ph hep-th

Hierarchical Axiverse

Hai-Jun Li This is my paper
classification hep-ph hep-th
keywords axiverseaxion-like particlesQCD axionmass mixingdecay constantsstochastic mixingmass hierarchySwampland constraints
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

The paper tries to show that the seemingly arbitrary spectrum of O(100) axion-like particles from string compactifications is actually structured by the mixing dynamics itself. It argues that QCD-induced mass mixing can only be well defined if no two ALPs are degenerate and if each mass crossing is individually resolvable, which forces adjacent ALP masses to be separated by roughly a factor of two. It further argues that stochastic mixing—in which ALPs mix regardless of their decay-constant ratios—drives the ratio f_Ai/f_a0 into two populations, one well below and one well above unity, leaving 0.1 ≲ R_f ≲ 10 empty. If these claims hold, the axiverse hierarchy is not an accidental input from string geometry or instanton physics but a consistency condition of low-energy mixing, and the empty window becomes a concrete falsifiable target for axion searches. The paper also verifies that the resulting spectrum passes the axion Weak Gravity Conjecture and Festina Lente bounds in standard large-volume compactifications.

Core claim

The central claim is that the hierarchical ordering of the string axiverse—in both masses and decay constants—emerges from the requirement of well-defined mixing. For the masses, non-degeneracy plus the condition that each temperature-driven QCD mass crossing be individually resolvable enforces m_Ai/m_Ai+1 ≲ 1/2, with the lightest ALP pushed down roughly as m_A1/m_a0 ≲ (1/2)^N. For the decay constants, stochastic mixing with binary labels d_i (recording only whether each ALP lies far below or far above f_a0) produces a distribution with R_f ≪ 1 and R_f ≫ 1 populated and the band 0.1 ≲ R_f ≲ 10 empty. The paper terms this emergent pattern the hierarchical axiverse and claims both patterns are

What carries the argument

The load-bearing object is the QCD-induced mass-mixing matrix built from the domain-wall-number matrix n_ij (Eq. 5), whose entries carry binary labels d_i = 0 or 1 depending on whether each ALP decay constant satisfies f_Ai ≪ f_a0 or f_Ai ≫ f_a0. Stochastic mixing lets every ALP participate in effective mixing regardless of these ratios, unlike the older maximal-mixing limit. The mass spacing factor 1/2 comes from requiring each crossing of the temperature-dependent QCD axion mass with an ALP mass to be individually resolvable, while the same binary structure in n_ij drives the two-population decay-constant distribution and the empty window.

Load-bearing premise

The load-bearing premise is that no ALP decay constant sits in the comparable range near the QCD axion’s decay constant; if one did, the stochastic-mixing construction—and the predicted empty window—would not go through.

What would settle it

An axion-search experiment that discovers an ALP whose decay-constant ratio R_f = f_Ai/f_a0 lies inside 0.1–10 would fill the empty window the paper predicts, directly contradicting the claim that mixing alone organizes the decay-constant distribution.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • The ALP mass spectrum in a string axiverse is a geometric ladder with adjacent states roughly a factor of two apart, instead of a free distribution.
  • Axion searches should find ALPs whose couplings correspond to R_f below about 0.1 or above about 10, with the intermediate band essentially empty unless additional UV physics contributes.
  • Larger axiverse populations push the lightest ALP mass down exponentially, making isocurvature constraints from many light fields more restrictive as N grows.
  • The hierarchical spectrum remains within the axion Weak Gravity Conjecture and Festina Lente windows in large-volume compactifications, so it is not ruled out by those quantum-gravity consistency conditions.
  • The empty window is a falsifiable prediction: an ALP discovered with R_f inside 0.1–10 would point to UV structure beyond mixing.

Where Pith is reading between the lines

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

  • If the mixing-origin hierarchy is correct, statistical distributions of axion spectra from random compactification ensembles could be re-examined for a depletion near R_f ~ 1; such a depletion would be a nontrivial cross-check of the mechanism.
  • The factor-of-two mass ladder resembles the spacing of random matrices with well-separated level crossings; a minimal random-matrix model might reproduce the same 1/2 factor and give an independent derivation.
  • Dedicated experimental coverage of the R_f band 0.1–10—through haloscope, helioscope, and astrophysical coupling limits—would provide a direct test of whether the gap holds.
  • The mixing-induced hierarchy could coexist with an instanton-induced hierarchy; distinguishing the two would require measuring correlations between adjacent mass spacings and decay-constant ratios.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The manuscript argues that the requirement of well-defined, non-degenerate QCD-induced mass mixing in a multi-ALP string axiverse forces the ALP masses into a hierarchical ordering with adjacent ratio m_Ai/m_A(i+1) ≲ 1/2, and that 'stochastic mixing' organizes the ALP decay constants into two populations separated by an essentially empty window 0.1 ≲ R_f ≲ 10. The claimed pattern is presented as UV-insensitive and as satisfying axion Weak Gravity Conjecture and Festina Lente bounds in large-volume compactifications. Section 2 introduces kinetic mixing and defines the physical basis; Section 3 defines the domain-wall-number matrix in Eq. (5), asserts the mass factor 1/2 from an unshown resolvability condition, and displays illustrative distributions in Figs. 1 and 2. Section 4 checks Swampland constraints. The Letter thus proposes a novel and falsifiable phenomenological scenario, but the central derivation is missing and the decay-constant gap appears to be built into the definitions rather than derived.

Significance. The question of whether low-energy mixing alone can impose a hierarchical structure on the axiverse is legitimate and potentially impactful, and the empty-window prediction is crisp and testable. The paper also correctly notes that Swampland bounds constrain f_i S_i rather than R_f, and the large-volume estimate saturating the WGC bound is a useful consistency check. However, the manuscript as submitted does not establish its central claims. The mass-ratio bound rests on an assertion with no calculation, and the two-population decay-constant structure is an input of Eq. (5) and the accompanying technical prerequisite, not an output of mixing dynamics. A convincing version of this work would require explicit diagonalization of the temperature-dependent mass matrix and a stochastic-mixing calculation that starts from a continuous prior over decay constants and demonstrates the emergence of bimodality. In its current form, the paper presents a scenario rather than a derivation.

major comments (4)
  1. [§3 (Hierarchical axiverse, mass distribution)] The central mass-spacing claim is asserted, not derived. The text states that 'we obtain the heavy and light eigenmasses m_{hi,li}' and solve (m_hi − m_li)|_{T→0} = (m_hi − m_li)|_{T→∞} to derive the mass factor 1/2, but no mass matrix, no explicit expressions for m_hi and m_li, and no solution are shown. This condition is the only quantitative input for the claimed bounds m_Ai/m_A(i+1) ≲ 1/2 and m_Ai/m_a0,0 ≲ (1/2)^{N−i+1}. Without the derivation, the mass hierarchy is unsupported; I could not verify the factor 1/2 or the stated dependence on f_Ai/f_a0.
  2. [§3, Eq. (5) and preceding paragraph] The decay-constant gap is imported, not emergent. The text imposes 'as a technical prerequisite, each ALP decay constant should lie outside the comparable regime f_Ai ∼ f_a0' and then defines d_i = 0 iff f_Ai ≪ f_a0 and d_i = 1 iff f_Ai ≫ f_a0 in Eq. (5). With d_i taking only these two values, every ALP has R_f either far below or far above unity before any mixing calculation is performed. The empty window 0.1 ≲ R_f ≲ 10 in Fig. 2 is therefore guaranteed by construction; the boundaries 0.1 and 10 are not derived from any equation. No calculation is shown in which a continuous prior over f_Ai evolves under mixing into a bimodal R_f distribution. Thus the claim that stochastic mixing 'drives' the decay constants into two populations is a restatement of Eq. (5), not a result.
  3. [§3, stochastic mixing and Fig. 2] The stochastic-mixing mechanism is never formulated in this paper. The text refers to Ref. [33] and contrasts stochastic with maximal mixing, but the only place the mixing structure enters is Eq. (5), which is a bookkeeping device based on the d_i labels. There is no Hamiltonian, no transition amplitude, no ensemble average, and no criterion that would make the two-population structure 'generic'. Consequently, the statement 'stochastic mixing generically drives R_f either well below or well above unity' has no computable content in this manuscript. Either the mechanism must be reproduced here or the claim must be deferred to a detailed companion paper.
  4. [§3, paragraph on non-degeneracy] The paper states that the two requirements (non-degeneracy and resolvability) 'force the spectrum to be hierarchically spaced'. The non-degeneracy condition only forbids exact equality; it does not by itself imply a factor-1/2 spacing. The spacing comes entirely from the unshown resolvability condition. The wording conflates a mild consistency condition with the strong quantitative hierarchy that is the paper's central result.
minor comments (5)
  1. [Fig. 2] The caption and text do not state how the plotted R_f values were generated. If they are random draws respecting the d_i labels, the figure is illustrative rather than evidence; the paper should say so explicitly.
  2. [Introduction / §3] The phrase 'stochastic mixing' is used as a term of art but defined only by citation to [33]; the Letter should include at least its defining equation or state clearly that it is a review of known results.
  3. [Abstract] The abstract's phrase 'well-defined QCD-induced mass mixing forces...' overstates the strength of the argument; 'we argue' or 'we conjecture' would be more accurate until the derivation is shown.
  4. [§4, Eq. (8)] Eq. (8) writes f_i S_i ∼ M_Pl ≲ M_Pl; the mixed '∼ ... ≲' notation is confusing. The paper should also clarify whether the estimate f_i ∼ M_Pl/V^{2/3}, S_i ∼ V^{2/3} holds in the same compactification that hosts the R_f ≫ 1 population.
  5. [References] Ref. [32] is cited as also obtaining a hierarchical mass spectrum; the distinction between the two origins would be clearer if the comparison quantified the statistical distribution in [32].

Circularity Check

3 steps flagged

The empty R_f window is imported, not emergent: Eq. (5)'s binary d_i labels and the stated 'technical prerequisite' f_Ai not ~ f_a0 force R_f away from unity before mixing is applied.

specific steps
  1. self definitional [Section 'Axiverse mass mixing', Eq. (5); Section 'Hierarchical axiverse', decay-constant paragraph and Fig. 2]
    "As a technical prerequisite, each ALP decay constant should lie outside the comparable regime f_Ai ∼ f_a0 ... The domain wall number matrix n_ij is given by n_ij = δ_ij + d_j δ_i0(1 − δ_j0) + (1 − d_i) δ_j0(1 − δ_i0), where δ_ij is the Kronecker delta, d_i = 0 iff A_i ≪ f_a0, and d_i = 1 iff A_i ≫ f_a0."

    Eq. (5) encodes the 'technical prerequisite' as binary labels d_i, forcing every ALP to have f_Ai ≪ f_a0 or f_Ai ≫ f_a0 by construction. The claimed emergent gap 0.1 ≲ R_f ≲ 10 in Fig. 2 is exactly the excluded 'comparable regime' f_Ai ∼ f_a0. No calculation starts from a continuous prior over f_Ai and produces a bimodal R_f distribution; the bimodality is an input, not an output of stochastic mixing dynamics.

  2. self citation load bearing [Section 'Axiverse mass mixing', paragraph beginning 'Note that we adopt a stochastic mixing mechanism...']
    "Note that we adopt a stochastic mixing mechanism, which allows effective mixing to occur across all N ALPs regardless of the relative magnitudes of their decay constants [33]. This cannot be realized in the previous maximal mixing mechanism..."

    The central mechanism that is claimed to organize the decay constants into two populations is not derived in this paper; it is imported from the author's own Ref. [33] (H.-J. Li, arXiv:2604.27784). The 'emergent, UV-insensitive' R_f gap therefore rests on a self-citation whose content is not reproduced or independently checked here. The paper's Conclusion attributes the gap to 'mixing itself,' but the mixing mechanism that does the work is exactly the cited prior work.

  3. other [Section 'Hierarchical axiverse', mass-distribution paragraph]
    "From the temperature-dependent mass matrix, we obtain the heavy and light eigenmasses m_{h,l}^i and solve (m_h^i − m_l^i)|_{T→0} = (m_h^i − m_l^i)|_{T→+∞} to derive the mass factor 1/2. For each mass crossing to be individually resolvable — i.e., for the crossings not to overlap — adjacent masses should satisfy m_Ai/m_Ai+1 ≲ 1/2 ..."

    The claimed derivation of the 1/2 mass spacing is not shown: no mass matrix, eigenmass expressions, or solution of the stated equation is exhibited. The 'individually resolvable' condition is itself a separation condition of roughly a factor of two, and the paper immediately identifies the threshold 1/2 as the condition's consequence. Thus the hierarchical mass-ordering m_Ai/m_Ai+1 ≲ 1/2 is effectively introduced as the definition of resolvability rather than derived from mixing dynamics.

full rationale

The paper's central decay-constant claim—that stochastic mixing generically produces two ALP populations separated by an empty window 0.1 ≲ R_f ≲ 10—is guaranteed by construction. Immediately before Eq. (5), the text imposes the technical prerequisite f_Ai not in the comparable regime f_Ai ∼ f_a0, and Eq. (5) encodes this as binary labels d_i = 0 or 1. With d_i taking only these values, every ALP has R_f either far below or far above unity, so the gap in Fig. 2 is an input, not an emergent prediction. The stochastic-mixing mechanism that is supposed to drive this pattern is imported wholesale from the author's own Ref. [33], and the companion mass hierarchy's 1/2 spacing is asserted through an unshown resolvability condition rather than a displayed calculation. The Swampland consistency checks are independent but peripheral to the hierarchical pattern claims. Overall, the main 'emergent, UV-insensitive' patterns reduce, at least partially, to the paper's own definitions and self-citations, warranting a score of 8.

Axiom & Free-Parameter Ledger

2 free parameters · 7 axioms · 0 invented entities

The paper introduces no new particles or forces, but it rests on several hand-selected modeling choices: the binary d_i classification in the domain wall matrix, the technical prerequisite excluding f_Ai ∼ f_a0, the imposed non-degeneracy, and the unshown resolvability condition. The Swampland bounds and QCD potential are standard external input.

free parameters (2)
  • Mass ratio threshold 1/2 = 0.5
    The value 1/2 is introduced as the reference spacing m_Ai/m_Ai+1 ≲ 1/2, said to follow from a solvability condition that is not shown. The bound tightens or loosens with f_Ai/f_a0, so it is a hand-chosen threshold rather than a derived constant.
  • Decay constant gap boundaries = 0.1 and 10
    The empty window 0.1 ≲ R_f ≲ 10 is defined as the 'comparable regime' excluded by the technical prerequisite. The boundary numbers are asserted without derivation or calibration.
axioms (7)
  • domain assumption QCD axion potential V_QCD = m_a0^2 f_a0^2 [1 - cos(theta_QCD)] with temperature-dependent mass (Eq. 3)
    Standard low-energy QCD axion potential, invoked in Eqs. (2)-(3).
  • domain assumption ALP masses are temperature-independent
    Assumed in the mass mixing Lagrangian (Eq. 4).
  • ad hoc to paper Domain wall number matrix n_ij = δ_ij + d_j δ_i0(1-δ_j0) + (1-d_i)δ_j0(1-δ_i0), with d_i = 0 iff f_Ai ≪ f_a0 and d_i = 1 iff f_Ai ≫ f_a0 (Eq. 5)
    This matrix encodes the two-population split of decay constants before it is claimed to emerge; it is central to the stochastic mixing mechanism.
  • ad hoc to paper Technical prerequisite: each ALP decay constant lies outside the comparable regime f_Ai ∼ f_a0
    Stated as essential for N≥2; directly implies the empty window 0.1 ≲ R_f ≲ 10.
  • ad hoc to paper No two ALPs may be exactly degenerate in mass
    Imposed as a requirement for well-defined mixing; this requirement is the stated engine of the mass hierarchy.
  • ad hoc to paper Resolvability condition (m_hi - m_li)|T→0 = (m_hi - m_li)|T→∞ leading to the factor 1/2
    The step that produces the mass spacing is not shown; the condition is stated without derivation.
  • domain assumption Stochastic mixing mechanism of Ref. [33] (same author)
    The mechanism that allows effective mixing across all N ALPs is taken from the author's own prior work; it does the work of generating the two populations.

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Hierarchical Axiverse." pith.science (2026). https://pith.science/paper/STFXTXC7

@misc{pith2026260803678,
  author       = {Pith},
  title        = {Pith review of: Hierarchical Axiverse},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/STFXTXC7}},
  note         = {Machine review of arXiv:2608.03678}
}
Share X Bluesky LinkedIn Reddit HN
abstract

String compactifications generically produce ${\cal O}(100)$ light axion-like particles, yet low-energy mixing has not previously been exploited as a structuring constraint on their spectrum. We show that the requirement of well-defined QCD-induced mass mixing forces the axion masses into a hierarchically spaced ordering, while stochastic mixing organizes the decay constants into a hierarchically split distribution -- two populations separated by an essentially empty window. Both patterns are insensitive to ultraviolet input and satisfy the relevant Swampland constraints within standard large-volume compactifications. We term this emergent pattern the hierarchical axiverse.

Figures

Figures reproduced from arXiv: 2608.03678 by Hai-Jun Li.

Figure 1
Figure 1. Figure 1: FIG. 1. Hierarchical axion mass distribution in the axiverse with [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Hierarchical axion decay constant distribution in the axiverse with [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

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

40 extracted references · 3 canonical work pages · 1 internal anchor

  1. [1]

    Witten, Phys

    E. Witten, Phys. Lett. B149, 351 (1984)

    1984

  2. [2]

    Choi, Phys

    K.-w. Choi, Phys. Rev. Lett.92, 101602 (2004), arXiv:hep-ph/0308024

    Pith/arXiv arXiv 2004

  3. [3]

    Svrcek and E

    P. Svrcek and E. Witten, JHEP06, 051 (2006), arXiv:hep-th/0605206

    Pith/arXiv arXiv 2006

  4. [4]

    Arvanitaki, S

    A. Arvanitaki, S. Dimopoulos, S. Dubovsky, N. Kaloper, and J. March-Russell, Phys. Rev. D81, 123530 (2010), arXiv:0905.4720 [hep-th]

    Pith/arXiv arXiv 2010

  5. [5]

    Cicoli, M

    M. Cicoli, M. Goodsell, and A. Ringwald, JHEP10, 146 (2012), arXiv:1206.0819 [hep-th]

    Pith/arXiv arXiv 2012

  6. [6]

    Demirtas, C

    M. Demirtas, C. Long, L. McAllister, and M. Stillman, JHEP04, 138 (2020), arXiv:1808.01282 [hep-th]

    Pith/arXiv arXiv 2020

  7. [7]

    Reig, JHEP09, 207 (2021), arXiv:2104.09923 [hep- ph]

    M. Reig, JHEP09, 207 (2021), arXiv:2104.09923 [hep- ph]

    Pith/arXiv arXiv 2021

  8. [8]

    Demirtas, N

    M. Demirtas, N. Gendler, C. Long, L. McAllister, and J. Moritz, JHEP06, 092 (2023), arXiv:2112.04503 [hep- th]

    Pith/arXiv arXiv 2023

  9. [9]

    J. M. Leedom, M. Putti, and A. Westphal, JHEP05, 271 (2026), arXiv:2509.03578 [hep-th]

    Pith/arXiv arXiv 2026

  10. [10]

    S. V. P. Fallon, J. Halverson, L. McAllister, and Y. Zhu, (2025), arXiv:2511.20458 [hep-th]

    arXiv 2025

  11. [11]

    Peccei and H

    R. Peccei and H. R. Quinn, Phys. Rev. Lett.38, 1440 (1977)

    work page 1977

  12. [12]

    Peccei and H

    R. Peccei and H. R. Quinn, Phys. Rev. D16, 1791 (1977)

    work page 1977

  13. [13]

    Weinberg, Phys

    S. Weinberg, Phys. Rev. Lett.40, 223 (1978)

    1978

  14. [14]

    Wilczek, Phys

    F. Wilczek, Phys. Rev. Lett.40, 279 (1978)

    1978

  15. [15]

    Preskill, M

    J. Preskill, M. B. Wise, and F. Wilczek, Phys. Lett. B 120, 127 (1983)

    1983

  16. [16]

    Abbott and P

    L. Abbott and P. Sikivie, Phys. Lett. B120, 133 (1983)

    1983

  17. [17]

    Dine and W

    M. Dine and W. Fischler, Phys. Lett. B120, 137 (1983)

    1983

  18. [18]

    D. J. E. Marsh, Phys. Rept.643, 1 (2016), arXiv:1510.07633 [astro-ph.CO]

    Pith/arXiv arXiv 2016

  19. [19]

    C. A. J. O’Hare, PoSCOSMICWISPers, 040 (2024), arXiv:2403.17697 [hep-ph]

    Pith/arXiv arXiv 2024

  20. [20]

    Daido, N

    R. Daido, N. Kitajima, and F. Takahashi, Phys. Rev. D 93, 075027 (2016), arXiv:1510.06675 [hep-ph]

    Pith/arXiv arXiv 2016

  21. [21]

    Li, Y.-Q

    H.-J. Li, Y.-Q. Peng, W. Chao, and Y.-F. Zhou, Phys. Lett. B849, 138444 (2024), arXiv:2310.02126 [hep-ph]

    Pith/arXiv arXiv 2024

  22. [22]

    Cyncynates and J

    D. Cyncynates and J. O. Thompson, Phys. Rev. D108, L091703 (2023), arXiv:2306.04678 [hep-ph]

    Pith/arXiv arXiv 2023

  23. [23]

    Li and Y.-F

    H.-J. Li and Y.-F. Zhou, Nucl. Phys. B1029, 117574 (2026), arXiv:2504.10170 [hep-th]

    Pith/arXiv arXiv 2026

  24. [24]

    Daido, N

    R. Daido, N. Kitajima, and F. Takahashi, Phys. Rev. D 92, 063512 (2015), arXiv:1505.07670 [hep-ph]

    Pith/arXiv arXiv 2015

  25. [25]

    S.-Y. Ho, K. Saikawa, and F. Takahashi, JCAP10, 042 (2018), arXiv:1806.09551 [hep-ph]

    Pith/arXiv arXiv 2018

  26. [26]

    Cyncynates, O

    D. Cyncynates, O. Simon, J. O. Thompson, and Z. J. Weiner, Phys. Rev. D106, 083503 (2022), arXiv:2208.05501 [hep-ph]

    Pith/arXiv arXiv 2022

  27. [27]

    Kitajima, J

    N. Kitajima, J. Lee, K. Murai, F. Takahashi, and W. Yin, Phys. Lett. B851, 138586 (2024), arXiv:2306.17146 [hep-ph]

    Pith/arXiv arXiv 2024

  28. [28]

    J. Lee, K. Murai, F. Takahashi, and W. Yin, JCAP10, 038 (2024), arXiv:2407.09478 [hep-ph]

    Pith/arXiv arXiv 2024

  29. [29]

    Murai, Y

    K. Murai, Y. Narita, F. Takahashi, and W. Yin, JHEP 04, 124 (2025), arXiv:2412.10232 [hep-ph]

    Pith/arXiv arXiv 2025

  30. [30]

    H.-J. Li, W. Chao, H.-K. Guo, and Y.-F. Zhou, Nucl. Phys. B1026, 117455 (2026), arXiv:2508.12722 [hep-ph]

    Pith/arXiv arXiv 2026

  31. [31]

    Murai, F

    K. Murai, F. Takahashi, and K. Tomiyama, JHEP07, 171 (2026), arXiv:2509.11032 [hep-ph]

    Pith/arXiv arXiv 2026

  32. [32]

    Baryakhtar, D

    M. Baryakhtar, D. Cyncynates, and E. Henry, (2026), arXiv:2602.23424 [hep-ph]

    Pith/arXiv arXiv 2026

  33. [33]

    Stochastic Axion Mixing: A General Mechanism Beyond Decay Constant Constraints

    H.-J. Li, (2026), arXiv:2604.27784 [hep-ph]

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  34. [34]

    Vafa, (2005), arXiv:hep-th/0509212

    C. Vafa, (2005), arXiv:hep-th/0509212

    Pith/arXiv arXiv 2005

  35. [35]

    Arkani-Hamed, L

    N. Arkani-Hamed, L. Motl, A. Nicolis, and C. Vafa, JHEP06, 060 (2007), arXiv:hep-th/0601001

    Pith/arXiv arXiv 2007

  36. [36]

    Harlow, B

    D. Harlow, B. Heidenreich, M. Reece, and T. Rudelius, Rev. Mod. Phys.95, 035003 (2023), arXiv:2201.08380 [hep-th]

    Pith/arXiv arXiv 2023

  37. [37]

    Montero, T

    M. Montero, T. Van Riet, and V. Venken, JHEP01, 039 (2020), arXiv:1910.01648 [hep-th]

    Pith/arXiv arXiv 2020

  38. [38]

    Montero, C

    M. Montero, C. Vafa, T. Van Riet, and V. Venken, JHEP 10, 009 (2021), arXiv:2106.07650 [hep-th]

    Pith/arXiv arXiv 2021

  39. [39]

    Guidetti, N

    V. Guidetti, N. Righi, V. Venken, and A. Westphal, JHEP01, 114 (2023), arXiv:2206.03494 [hep-th]

    Pith/arXiv arXiv 2023

  40. [40]

    Li, Phys

    H.-J. Li, Phys. Lett. B873, 140141 (2026), arXiv:2510.22569 [hep-ph]

    arXiv 2026

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.