{"id":"e1903885-a0f6-4aed-aecc-e80398e92104","arxiv_id":"2607.23603","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Force-monopole stiffnesses in a mean-field glass model obey an exact identity with global susceptibilities, reconstruct the vibrational spectrum, and define a stabilized-sector scale that tracks the boson-peak frequency in the non-equilibrium RSB phase.","lead":"A mean-field glass model is probed with local force-monopole responses, yielding an exact stiffness–susceptibility identity and a characteristic soft-mode scale tied to the boson peak. The work links non-equilibrium replica-symmetry breaking to an effective random-potential picture and to micromechanics used in computer glasses.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"Quench-induced diagonal–coupling correlations remain the key unproven input to the κ_i identity; current checks test its consequences, but not the conditional cavity average directly.","rationale":"This is the same load-bearing assumption identified by the Reader, stated more precisely as a conditional cavity-averaging issue rather than generic diagonal–off-diagonal independence. It is an internal approximation, not merely a departure from consensus.\n\nThe paper nevertheless supplies unusually strong consequential checks: the predicted anomalous \\(N^{-1/5}\\) error scaling in Fig. A1, the pairwise and global susceptibility identities in Appendix B, the full VDoS reconstruction in Fig. 3, and a useful spherical 3-spin negative control. Those results make an \\(O(1)\\) failure unlikely at the tested parameters, particularly because several checks extend to \\(J=0.01\\). They do not completely prove asymptotic conditional self-averaging, but they are direct enough that the concern should not overturn the verdict without a failed cavity-residual test.\n\nAccordingly, I would retain ACCEPT with medium correctness risk. A persistent conditional bias would specifically undermine the pointwise map from \\(\\kappa_i\\) to \\(a_i\\), and thereby the reconstruction and stabilized-sector claims, rather than merely the broader effective-potential analogy.","tokens_in":28003,"tokens_out":7036,"duration_ms":225811,"concrete_test":"For quenched minima at \\(J=0.01,0.05,0.5\\) and \\(N=1024\\)–\\(32000\\), compute each cavity resolvent \\(G^{(i)}(0)\\), row \\(J_i\\), and residual \\(R_i=J_i^T G^{(i)}(0)J_i+J^2\\chi\\). Bin \\(R_i\\) by \\(a_i=k_i+x_i^2/2\\) and estimate the conditional bias \\(\\mathbb E[R_i\\mid a_i]\\), normalized by the local predicted scale. If this bias does not decay with \\(N\\), Eq. (A2) fails conditionally; if it vanishes, the concern does not land.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equation (A2) replaces the cavity quadratic form \\(Q_i=\\sum_{j,k\\ne i}J_{ij}G^{(i)}_{jk}J_{ki}\\) by \\((J^2/N)\\operatorname{Tr}G^{(i)}\\). For a fresh GOE row independent of the cavity, this is the standard result. Here, however, the quenched minimum makes \\(a_i=k_i+x_i^2/2\\) a functional of the full coupling matrix, including row \\(i\\), while \\(G^{(i)}\\) depends on the remaining couplings. Thus the step required for Eq. (6), and especially for the subsequent pointwise \\(a_i\\leftrightarrow\\kappa_i\\) map, is not merely an unbiased global average of \\(Q_i\\), but conditional self-averaging of \\(Q_i\\) given \\(a_i\\).\n\nIf conditioning on \\(a_i\\) induces an \\(O(1)\\) bias in \\(Q_i+J^2\\chi\\), or a bias that grows toward small \\(J\\) where \\(\\chi_{SG}\\sim J^{-2}\\), Eq. (6) could remain accurate for selected global moments while failing as a per-oscillator distributional identity. That would weaken the spectrum reconstruction and the stabilized-sector interpretation. The finite-size test in Fig. A1 and the global-identity and spectrum checks substantially mitigate this concern, but they do not isolate the conditional cavity covariance itself.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript studies a fully-connected mean-field glass Hamiltonian (Eq. 1) in its non-equilibrium RSB phase (instantaneous quench, h=0, gapless p(k)). It defines a local micromechanical observable — the force-monopole stiffness kappa_i = (M^{-1})_{ii}/(M^{-2})_{ii} — as the mean-field analog of the dipole (\"pinching\") stiffness in finite-dimensional glasses. The central analytic result is the mean-field identity Eq. (6), kappa_i = [(k_i - J^2 chi) + x_i^2/2]/(1 + J^2 chi_SG), derived via Schur complement and GOE averaging, with anomalous finite-N fluctuations predicted to scale as N^{-1/5} (Appendix A, Fig. A1). Corollaries include pairwise and global estimators of chi and chi_SG from kappa statistics alone (Appendix B), a reconstruction of the full VDoS from {kappa_i} validated against direct diagonalization (Fig. 3a), the identification of a \"stabilized sector\" {k_i < J^2 chi} whose conditional stiffness distribution collapses and whose mean scales as <kappa>_st ~ omega_p^2 (Figs. 4-5), a mesoscopic-vs-global modulus comparison (Fig. 6), and an extended correspondence with the equilibrium effective random potential, including chi_SG ~ J^{-2} (Fig. 7) and omega_p ~ (Delta f)^{1/3} (Fig. 8). Appendix H shows the monopole response is ill-defined in the spherical 3-spin model, supporting the claimed link to the quartic VDoS tail.","tokens_in":28330,"tokens_out":5044,"duration_ms":197401,"significance":"If the identity holds as argued, the paper delivers several genuinely strong results: a parameter-free, asymptotically exact relation between a local response and global susceptibilities; an explicit, testable prediction for anomalous finite-size fluctuations (N^{-1/5}) that is numerically confirmed; a reconstruction of the full VDoS from monopole statistics validated against direct diagonalization on identical disorder realizations (Fig. 3a); a falsifiable scaling bridge (Eq. 9) between mesoscopic stiffness and the boson-peak frequency that is in principle exportable to laboratory glasses; and a clean negative result (Appendix H) delineating when the micromechanical description fails. The work meaningfully clarifies the structure of this mean-field glass model and strengthens its contact with finite-dimensional glass physics.","major_comments":[{"comment":"The derivation replaces the cavity quadratic form Q_i = sum_{j,k != i} J_ij G^{(i)}_{jk} J_ki by (J^2/N) Tr G^{(i)}, assuming no correlation between the diagonal a_i = k_i + x_i^2/2 and the off-diagonal row {J_ij}. At a quenched minimum, however, x_i is a functional of the full coupling matrix including row i, so the assumption is not exact; what Eq. (6) actually requires is conditional self-averaging of Q_i given a_i. The presented checks (Fig. A1 error distribution, Figs. 3a, A3) test consequences of the identity but average over all oscillators. Since the stabilized-sector construction (Sect. VI) and the a_i <-> kappa_i map (Eq. C2) use the identity pointwise, I ask for a direct test of structured bias, e.g., the conditional mean of Delta kappa_i (Eq. A6) versus k_i, or restricted to k_i < J^2 chi. If no bias is found, a short statement to that effect suffices.","section":"Appendix A, Eq. (A2)"},{"comment":"The finite-N convergence test of the identity is shown only at J = 0.5. Because chi_SG ~ J^{-2} (Fig. 7), the fluctuation terms O(sqrt(chi_SG/N)) in Eqs. (A3)-(A4) and the extreme-value scale omega_min both worsen as J decreases, while the headline scaling results (Figs. 3-5, 8) extend to J = 0.01 at fixed N = 8192. A Fig. A1-style convergence check at a small J (e.g., 0.05), or N-dependence of the data in Fig. 5, is needed to establish that the N^{-1/5} fluctuation regime, rather than a J-dependent systematic error, controls the small-J points that carry the scaling claims.","section":"Appendix A, Eqs. (A3)-(A7) and Fig. A1"},{"comment":"The claimed correspondence between the monopole identity and the effective potential (that Eqs. (6) and (11) are 'essentially identical') rests on J^2 chi_SG being J-independent, supported only by the numerical line chi_SG = J^{-2} in Fig. 7. The precision, the J-range over which it holds, and its status (exact identity vs. approximate scaling) are not quantified, and no analytic argument (e.g., from marginal stability) is offered. Since this is a central conceptual claim of Sects. VIII-IX, the authors should quantify deviations from chi_SG ~ J^{-2}, state explicitly whether they conjecture it to be exact, and temper the wording if it is only approximate.","section":"Sect. VIII, Fig. 7"}],"minor_comments":[{"comment":"The cutoff omega_g is defined operationally by A_g omega^4 / D_G(omega) >= 2 (Appendix D). The slope and quality of the omega_p ~ omega_g relation in Fig. 3b presumably depend on this threshold; please report sensitivity (e.g., thresholds 1.5 and 3) in an inset or supplementary figure.","section":"Appendix D; Fig. 3b"},{"comment":"The pair estimators in Eq. (B1) are singular as kappa_i -> kappa_j, and second moments are computed only from pairs with |kappa_i - kappa_j| > 10^{-2}. Please show that the N^{-1/5} scaling in Fig. A2 is robust to this cutoff and comment on whether p(Delta_chi^{(ij)}) has heavy tails.","section":"Appendix B, Eq. (B1) and Fig. A2"},{"comment":"The definition delta f_i = (k_i - J^2 chi) x_i + x_i^3/6 is asserted by analogy to v'_eff(x*) = 0. It would help to state explicitly that this follows from mechanical equilibrium, sum_j J_ij x_j = -(k_i x_i + x_i^3/6), so that delta f_i is minus the cavity field including the Onsager-like reaction term J^2 chi x_i.","section":"Sect. IX"},{"comment":"Fig. 1b is a schematic 2D embedding of a fully connected model; the caption should state even more explicitly that positions carry no meaning, and the participation number N_e of the response vector z_i should be defined in the main text, not only used.","section":"Sect. III, Fig. 1b"},{"comment":"Please state the convergence criterion and lambda-grid used when iterating Eq. (7) with eta = 10^{-8}, and clarify why Fig. A3 uses M = 200 realizations while M = 2000 is used elsewhere (Appendix G).","section":"Appendix C; Fig. A3"},{"comment":"Typos: 'arranging the the remaining oscillators' (Fig. 1 caption); 'the scaling of the the mean' (after Eq. A5); 'degree of freedoms' (abstract and Sect. II); 'mesocopic' (Sect. III); 'two subpopulation' (Sect. VI).","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript sits within a research program largely built by the authors themselves: the model [47–52] and the dipole-micromechanics framework [53–55] are both from the same group and its close collaborators. This is natural for a follow-up study and the derivations here are self-contained, but the editor may wish to secure one report from the replica/cavity spin-glass community specifically on Appendix A, since the independence assumption underlying Eq. (A2) is the one step that resembles, but is not identical to, the standard fresh-GOE-row argument used in the equilibrium analysis of [47]. The empirical relation χ_SG ∼ J^{-2} (Fig. 7), if exact, very likely has a simple marginal-stability origin; its analytic explanation would considerably strengthen the paper and might be worth encouraging regardless of the revision outcome."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The one thing worth knowing is that they finally give the fully-connected model a concrete local probe that mirrors the finite-D dipole “pinching” work, and they get an asymptotically exact identity out of it. κ_i = [(k_i − J²χ) + ½x_i²] / (1 + J²χ_SG) is derived by Schur + GOE averaging, the N^{-1/5} fluctuation scaling is predicted and checked, and the same {κ_i} rebuild the entire D_G(ω) to high accuracy against direct diagonalization. The stabilized-sector conditional mean then tracks ω_p² cleanly, and they close the loop with Δf^{1/3}. That package is new relative to their earlier spectral/yielding papers.\n\nWhat works: the math is standard RMT done carefully, the negative control (spherical 3-spin, ω² spectrum → ill-defined monopoles) is exactly the right sanity check, and the appendices actually deliver the finite-size, pairwise, and global corollaries instead of waving at them. The analogy to the RS effective potential is structural and scaling-based rather than a fresh replica solution, but they do not oversell it as more than that.\n\nThe soft spot the stress-test flags is real: the derivation assumes the cavity quadratic form self-averages conditionally on a_i. Quench-induced diagonal–coupling correlations could in principle bias the pointwise map even if global moments look fine. Their Fig. A1, spectrum reconstruction, and χ_SG ∼ J^{-2} checks make a large bias unlikely, but they never isolate the conditional covariance itself. That is a genuine caveat, not a deal-breaker; it is the usual mean-field leap and the numerics constrain it tightly. Operational definitions (ω_g threshold, pair cut) are minor. Heavy self-citation is expected for this series and not load-bearing for the new identities.\n\nThis is for people who already care about mean-field glasses, soft spots, or boson-peak scales. If you work on micromechanics or RSB spectra, read it; the reconstruction chain and stabilized-sector criterion are usable ideas. I would send it to referees without hesitation and would cite the identity and the ⟨κ⟩_st ∼ ω_p² result. Bring it to reading group if the group is in disordered systems or glasses.","headline":"Solid technical advance: monopole identity, VDoS-from-κ reconstruction, and stabilized-sector link to ω_p inside their non-eq RSB model; the cavity-independence assumption is the only real soft spot and is already stress-tested by consequences.","tokens_in":29193,"tokens_out":615,"would_cite":true,"duration_ms":19758,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Local force-monopole responses in a mean-field glass model obey an exact identity that recovers the vibrational spectrum and realizes the effective random potential deep in the non-equilibrium RSB phase.","keywords":["mean-field glasses","replica symmetry breaking","force monopole response","vibrational density of states","boson peak","spin-glass susceptibility","marginal stability","effective random potential"],"falsifier":"Compute monopole stiffnesses and global susceptibilities on large quenched samples and test whether the measured κ_i equal the right-hand side of the identity within the predicted N^{-1/5} fluctuations; systematic O(1) deviations at large N would falsify the claim.","tokens_in":28927,"feed_emoji":"␙","tokens_out":1141,"duration_ms":24500,"temperature":0.7,"pith_summary":"This paper studies a fully connected mean-field Hamiltonian for glasses after an instantaneous quench into its non-equilibrium replica-symmetry-breaking phase. It defines the linear response to a local force monopole on each oscillator and derives an asymptotically exact identity that ties each monopole stiffness to that oscillator's bare stiffness and displacement and to two global susceptibilities. From the statistics of those monopole stiffnesses alone one can reconstruct the full nonphononic vibrational density of states, including its universal quartic soft tail and its intrinsic boson-peak frequency. A stability criterion then isolates a stabilized sector of oscillators whose mean monopole stiffness scales as the square of that boson-peak frequency, showing that these local responses are the concrete collective degrees of freedom of an effective random potential previously derived only near equilibrium. The result supplies a direct bridge between mean-field glass theory and the micromechanical observables measured in finite-dimensional computer glasses.","feed_headline":"Local pinches recover a glass model's full vibration spectrum","feed_subtitle":"Monopole stiffnesses obey an exact identity and track the boson-peak scale deep in the non-equilibrium phase","key_machinery":"The mean-field monopole-stiffness identity κ_i = [(k_i − J²χ) + ½ x_i²] / (1 + J² χ_SG), obtained from the resolvent of the Hessian via the Schur complement and random-matrix self-consistency. It converts local response data into global susceptibilities, enables spectrum reconstruction, and defines the stabilized sector whose mean stiffness tracks the boson peak.","core_discovery":"In the non-equilibrium RSB phase of the mean-field glass model, the monopole stiffness satisfies the identity κ_i = [(k_i − J²χ) + ½ x_i²] / (1 + J² χ_SG). The vibrational density of states can be reconstructed from the set {κ_i} alone, and the conditional mean stiffness of the stabilized sector (k_i < J²χ) scales as ⟨κ⟩_st ∼ ω_p², furnishing a concrete micromechanical realization of the collective degrees of freedom of the effective random potential.","pith_inferences":["If the same monopole-to-spectrum map holds on sparse graphs, the identity would give a practical diagnostic of RSB-like marginality in finite-connectivity glass models.","Laboratory 'pinching' protocols that measure local stiffness distributions could test whether real glasses obey ⟨κ⟩_st ∼ ω_BP² without needing full vibrational spectroscopy.","The stabilized-sector oscillators are natural candidates for the soft spots that control yielding and mechanical memory under cyclic drive already studied in the same model."],"forward_implications":["The full nonphononic vibrational spectrum, including the quartic tail and boson-peak scale, is encoded in local monopole-response statistics alone.","Mesoscopic elasticity (stabilized-sector mean stiffness) varies more strongly with disorder than the global modulus, matching the pattern seen in computer glasses.","The effective random-potential description remains valid deep inside the non-equilibrium RSB phase even when no equilibrium RS phase exists.","Force-fluctuation scale Δf satisfies ω_p ∼ (Δf)^{1/3}, linking random forces in the effective potential to the boson peak.","Models with a softer ω² spectrum (divergent spin-glass susceptibility) cannot support a well-defined local monopole stiffness."],"fun_headline_variants":["Local monopoles rebuild full vibration spectrum in glass RSB phase","Monopole stiffness identity recovers glassy vibrational DOS","Local force pinches track boson-peak scale in non-equilibrium glass","Micromechanical responses reveal mean-field glass collective modes","Monopole statistics alone reconstruct the model's vibrational spectrum"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The derivation assumes that the diagonal Hessian entries remain statistically uncorrelated with the random off-diagonal couplings after the quench; if those correlations stay finite at large system size, the identity and the spectrum reconstruction both fail.","fun_headline_variants_meta":{"raw":{"variants":["Local monopoles rebuild full vibration spectrum in glass RSB phase","Monopole stiffness identity recovers glassy vibrational DOS","Local force pinches track boson-peak scale in non-equilibrium glass","Micromechanical responses reveal mean-field glass collective modes","Monopole statistics alone reconstruct the model's vibrational spectrum"]},"model":"grok-4.5","effort":"low","cost_usd":0.005438,"raw_usage":{"total_tokens":1528,"prompt_tokens":878,"num_sources_used":0,"completion_tokens":64,"cost_in_usd_ticks":54384000,"prompt_tokens_details":{"text_tokens":878,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":586,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":878,"tokens_out":64,"duration_ms":9412,"temperature":1.0,"reasoning_tokens":586,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T17:48:11.359000+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Compute monopole stiffnesses and global susceptibilities on large quenched samples and test whether the measured κ_i equal the right-hand side of the identity within the predicted N^{-1/5} fluctuations; systematic O(1) deviations at large N would falsify the claim.","supporting_citations":[],"review_version":1}