{"id":"73310365-7242-4172-b20e-3d84adf8f535","arxiv_id":"2607.07081","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":4,"one_line_summary":"GOE-like spectral chaos is neither necessary nor sufficient for quantum Mpemba crossings in a clean U(1)-conserving XXZ chain; the crossing is controlled by local charge-sector coherence structure instead.","lead":"This paper shows that chaotic energy-level statistics do not determine whether the quantum Mpemba effect occurs in a spin chain. The finding matters because it separates global spectral chaos from local symmetry-restoration dynamics, redirecting the search for the mechanism behind anomalous relaxation.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Finite-size scope is acknowledged but the detuned-staggered control's size-dependent wave vectors leave a narrow gap","rationale":"The reader correctly identifies the finite-size representativeness of the detuned-staggered control as the weakest assumption. The paper is transparent about this limitation and scopes its claims accordingly. The methodology is sound: symmetry-resolved level statistics, operationally defined crossing fractions, and the Frobenius coherence diagnostic provide three distinct layers of evidence. The block-resolved analysis at N=20 (Fig. S2) gives a concrete mechanistic explanation for why total coherence inversion does not guarantee entanglement-asymmetry crossing, which is the paper's most novel contribution. The concern that the detuned-staggered wave vector is size-dependent is real but does not undermine the stated finite-size claim. The verdict of ACCEPT with HIGH confidence is reasonable; I would not lower confidence below MEDIUM-HIGH because the key separation exists at only one parameter point per size, but the mechanistic analysis provides independent support that does not rely solely on the spectral-crossing coincidence.","tokens_in":14467,"tokens_out":2349,"duration_ms":59367,"concrete_test":"Fix q/π=1.028 and compute ⟨r⟩ and f_cross at N=16, 18, 20, 22 (if feasible). If ⟨r⟩ remains GOE-like (≳0.52) and f_cross remains 0 across these sizes, the finite-size concern is substantially weakened. If ⟨r⟩ drifts below 0.50 or f_cross becomes nonzero, the detuned-staggered control is size-tuned rather than robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central separation rests on two N=20 data points: generic field (GOE-like, crossings) and detuned-staggered field (GOE-like, no crossings). The detuned-staggered wave vector is not fixed across sizes (q/π = 1.050, 1.038, 1.040, 1.034, 1.028 for N=12–20), so there is no single Hamiltonian being scaled. The paper explicitly acknowledges this as a finite-size diagnosis, not a thermodynamic claim, which is appropriate. However, the claim's force depends on the detuned-staggered point remaining GOE-like at N=20 while having no crossings — and this is established for exactly one wave vector at one system size. If one fixed q/π=1.028 and increased N, the spectrum could drift toward or away from GOE, and the crossing behavior could change. The paper does not test this. This is not an internal inconsistency; it is a scope limitation that the authors mostly own, but the reader's verdict of HIGH confidence may be slightly generous given that the key negative control exists at a single size-dependent parameter point. The Frobenius coherence analysis (Fig. 4) strengthens the argument by providing a mechanistic layer independent of the spectral diagnostic, which partially compensates.","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"The manuscript investigates whether spectral chaos (as diagnosed by GOE-like level statistics) determines the occurrence of quantum Mpemba effect (QME) crossings in the entanglement asymmetry. Using a clean U(1)-conserving XXZ spin chain perturbed by longitudinal field textures, the authors compare adjacent-gap ratios with sustained-crossing fractions of the entanglement asymmetry for the same Hamiltonians. The central finding is a negative result: GOE-like spectra can occur with or without crossings, and crossings can appear away from the GOE reference. The authors further introduce a Frobenius charge-sector coherence diagnostic, showing that even inversion of total local charge-sector coherence does not guarantee an entanglement-asymmetry crossing, thereby isolating the crossing mechanism to entropy-weighted, block-resolved charge-sector structure.","tokens_in":14640,"tokens_out":1093,"duration_ms":94776,"significance":"The paper addresses a well-posed and timely question: whether chaotic thermalization is the organizing mechanism behind symmetry-restoration quantum Mpemba crossings. The negative result—that spectral chaos is neither necessary nor sufficient for crossings—is a useful and clarifying contribution. The diagnostic hierarchy (global level statistics, total Frobenius coherence, entropy-weighted asymmetry) is a clean conceptual framework. The numerical methodology is clearly specified: exact diagonalization, symmetry-resolved adjacent-gap ratios, sustained-crossing criteria with stated tolerances, and position-averaged entanglement asymmetry. The uniform N=20 point correctly resolves reflection and spin-flip parities before computing level statistics. The block-resolved decomposition (Fig. S2) provides a concrete mechanistic picture for the finite-size separation.","major_comments":[{"comment":"The central separation at N=20 rests on two GOE-like data points with opposite crossing responses: the generic field (q/π=0.5, ⟨r⟩≈0.556, f_cross=1) and the detuned-staggered field (q/π=1.028, ⟨r⟩≈0.532, f_cross=0). The detuned-staggered wave vectors are size-dependent (q/π=1.050, 1.038, 1.040, 1.034, 1.028 for N=12–20; Table S3), so there is no single Hamiltonian being scaled. The paper acknowledges this as a finite-size diagnosis, which is appropriate. However, the claim's force depends on the detuned-staggered point remaining GOE-like while having no crossings, and this is established for a size-dependent parameter point at each N. The authors should briefly discuss whether fixing q/π=1.028 and increasing N (or scanning q at fixed N=20) would preserve the separation, or at minimum state more explicitly that the near-staggered control is defined by its proximity to the staggered point,","section":null},{"comment":"The staggered field (q=π) at N=20 is reported with ⟨r⟩=0.4244 (Table S2), which is intermediate between Poisson (0.386) and GOE (0.531). The text states this value is the midpoint of two clean neighboring spectral windows (0.4439 and 0.4048) because the nominal central shift was numerically ill-conditioned. While the authors are transparent about this, the staggered point is one of only four N=20 data points and is the sole example of crossings away from GOE. A brief comment on the spectral classification of the staggered field (e.g., whether it is better described as partially chaotic or near-integrable) would strengthen the interpretation.","section":null}],"minor_comments":[{"comment":"In Fig. 2, the inset annotations (e.g., '5 70.77' and '0.91' in the generic panel) are not self-explanatory. A brief caption note clarifying what these numbers represent would help the reader.","section":null},{"comment":"The phase offset ϕ=0.37 is stated as fixed for all calculations. While the authors explain it is not fitted, a sentence noting that the results are insensitive to small changes in ϕ (or acknowledging that this has not been checked) would be useful.","section":null},{"comment":"In the Supplemental Material, Sec. S4, the quadratic expansion (Eq. S12) is stated to have median state-level relative errors of 10.2% and 8.0% with worst errors of about 27%. The authors correctly note it is used only for channel identification, not quantitative prediction. This caveat is clearly stated but could be slightly more prominent in the main text where Eq. (9) is discussed.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The reader's assessment of HIGH confidence is largely justified. The skeptic's concern about the size-dependent detuned wave vectors is valid but the authors are appropriately cautious, framing the result as a finite-size mechanism diagnosis. The Frobenius coherence analysis (Fig. 4) and block-resolved decomposition (Fig. S2) provide a mechanistic layer that partially compensates for the single-point nature of the N=20 detuned-staggered control. The paper is a solid negative-result contribution and the recommendation of minor revision reflects the desire for slightly more discussion of scope, not a fundamental deficiency."},"author_rebuttal":null,"desk_editor":{"model":"glm-5.2","letter":"The main thing to know: this paper shows that GOE-like level statistics neither guarantee nor prevent entanglement-asymmetry crossings in a clean U(1)-conserving XXZ chain. The key evidence is two N=20 points with GOE-like spectra but opposite crossing behavior—generic field crosses, detuned-staggered doesn't—and a staggered field that crosses despite being away from GOE. That's a clean negative result within one Hamiltonian family, which is genuinely useful for the subfield. A lot of prior work studied Mpemba crossings in integrable vs. chaotic systems separately; this directly compares the two diagnostics for the same Hamiltonians, and the separation is real where it matters most (N=20, GOE-like regime, opposite responses). The Frobenius coherence analysis is a nice addition: it shows total charge-sector coherence can invert without producing an entanglement-asymmetry crossing, which sharpens the message that entropy weighting matters, not just raw off-block-diagonal weight. The block-resolved decomposition at N=20 identifying edge vs. central sector competition is a reasonable mechanistic story for the finite-size data. The methodology is solid: symmetry-resolved level statistics, operationally defined crossing criterion with stated tolerances, position averaging, and the uniform N=20 point correctly resolves reflection and spin-flip parities before computing gap ratios. No fitted parameters are introduced to manufacture the separation. The soft spot is the finite-size scope, and the authors mostly own it. The detuned-staggered wave vectors are size-dependent (q/π ranges from 1.050 to 1.028 across N=12–20), so there's no single Hamiltonian being scaled. The key negative control—GOE-like spectrum, no crossings—exists at exactly one wave vector at N=20. If you fixed q/π=1.028 and pushed to larger N, the spectrum could drift and the crossing behavior could change. The paper doesn't test this. The stress-test note flags this correctly, but it's a scope limitation, not an internal inconsistency, and the Frobenius analysis partially compensates by providing a mechanistic layer independent of the spectral diagnostic. The reader's HIGH confidence is slightly generous but not unreasonable given that the central claim is a finite-size diagnosis explicitly framed as such. This is for people working on quantum thermalization, Mpemba effects, and symmetry restoration in spin chains. It deserves a serious referee who can assess whether the finite-size mechanism is convincing and whether the scope claims are appropriately bounded.","headline":"Clean finite-size separation of spectral chaos from Mpemba crossings; scope is honestly limited","tokens_in":15152,"tokens_out":596,"would_cite":true,"duration_ms":71263,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["05.45.Mt","03.67.Mn","75.10.Pq"],"model":"glm-5.2","headline":"Chaos Doesn't Decide When Quantum Memory Flips","keywords":[],"falsifier":"If at larger system sizes the detuned-staggered field develops entanglement-asymmetry crossings (or the staggered field loses them), the claimed separation between spectral chaos and Mpemba crossings would narrow or collapse.","tokens_in":14723,"feed_emoji":"🔀","tokens_out":967,"duration_ms":91872,"temperature":0.7,"pith_summary":"The paper asks a sharp question about the quantum Mpemba effect — the phenomenon where a state that starts further from symmetry can restore that symmetry faster than one that starts closer. The natural guess is that quantum chaos (in the technical sense of energy levels repelling each other like a random matrix, known as GOE statistics) is what drives this ordering reversal. The authors test this directly in a clean spin chain that conserves total magnetization, by comparing two things for the same Hamiltonian: whether the spectrum looks chaotic (GOE-like level statistics) and whether the entanglement asymmetry curves actually cross. They find no correlation. A Hamiltonian can have GOE-like chaos and still show no Mpemba crossing; it can show crossings while being far from the GOE reference. The paper then drills one level deeper: even the total amount of charge-sector coherence (the off-diagonal quantum correlations between different magnetization sectors of a subsystem) can invert its ordering without producing an entanglement-asymmetry crossing. The reason is that the entanglement asymmetry weights those coherence blocks by the eigenvalue spectrum of the symmetrized density matrix, so an unweighted sum inverting does not guarantee the entropy-weighted ordering inverts. The crossing is controlled by how local coherence is distributed across charge blocks and weighted, not by global spectral chaos.","feed_headline":"Chaos Doesn't Decide When Quantum Memory Flips","feed_subtitle":"In a clean spin chain, GOE-level chaos is neither required nor sufficient for the quantum Mpemba crossing — local coherence structure is.","key_machinery":"The central objects are: (1) the entanglement asymmetry ΔS_A, defined as the von Neumann entropy difference between the symmetrized reduced density matrix (where off-diagonal charge blocks are removed) and the full reduced density matrix — it quantifies how much U(1) symmetry breaking remains locally; (2) the adjacent-gap ratio ⟨r⟩, a standard measure of whether energy levels repel (GOE, chaotic, ⟨r⟩≈0.531) or cluster (Poisson, integrable, ⟨r⟩≈0.386); (3) the Frobenius charge-sector coherence C_A, an unweighted norm of the off-block-diagonal part of the reduced density matrix; and (4) the key identity ΔS_A = D(ρ_A ∥ ρ_A^sym), expressing entanglement asymmetry as a quantum relative entropy,  ","core_discovery":"Within a single clean U(1)-conserving XXZ spin chain family, changing only the spatial profile of a longitudinal magnetic field separates three distinct levels of information that had been assumed to be linked: (1) global spectral chaos measured by adjacent-gap ratio statistics, (2) the total local charge-sector coherence measured by a Frobenius norm, and (3) the entropy-weighted ordering measured by entanglement asymmetry. The key finding is that GOE-like level statistics are neither necessary nor sufficient for Mpemba crossings, and even inversion of the total charge-sector coherence is not sufficient — the crossing depends on block-resolved coherence weighted by the eigenvalue structure, ","pith_inferences":[],"forward_implications":["If spectral chaos does not control Mpemba crossings, then the search for Mpemba effects in experiments should focus on engineering field textures and initial-state coherence channels rather than on achieving chaotic spectra.","The separation between total coherence inversion and entropy-weighted inversion suggests a design principle: to induce or suppress Mpemba crossings, one should target specific charge-block coherence channels and their eigenvalue-dependent weights, not the overall coherence magnitude.","The block-resolved analysis showing edge-sector versus central-sector competition could generalize to other symmetry-restoration settings beyond U(1), wherever the reduced density matrix has a block structure that is differentially weighted by entropy."],"fun_headline_variants":["Quantum Mpemba Crossings Survive Without Spectral Chaos","Spectral Chaos Is Neither Required Nor Enough for Mpemba Crossings","Chaos Statistics and Quantum Memory Flips Decouple in a Spin Chain","What Controls Quantum Mpemba Crossings? Not Spectral Chaos","GOE Spectra Don't Decide Mpemba Crossings in Clean Spin Chains"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The entire separation is established at finite system sizes (12 to 20 spins), and the detuned-staggered wave vectors are adjusted per system size rather than following a fixed scaling trajectory, so whether the texture-selective classification persists in the thermodynamic limit remains open.","fun_headline_variants_meta":{"raw":{"variants":["Quantum Mpemba Crossings Survive Without Spectral Chaos","Spectral Chaos Is Neither Required Nor Enough for Mpemba Crossings","Chaos Statistics and Quantum Memory Flips Decouple in a Spin Chain","What Controls Quantum Mpemba Crossings? Not Spectral Chaos","GOE Spectra Don't Decide Mpemba Crossings in Clean Spin Chains","Local Coherence, Not Chaos, Gates the Quantum Mpemba Crossing","Three Supposed Links in Quantum Mpemba Physics Pull Apart","Quantum Memory Reversal Outruns Chaos in a U(1) Spin Chain","Entanglement Asymmetry Crossings Track Coherence, Not GOE Statistics","Charge-Sector Coherence Blocks Decide Mpemba Crossings, Not Chaos","Quantum Mpemba Crossings Appear With and Without Chaotic Spectra","Spectral Chaos and Mpemba Crossings: Linked in Theory, Split in Practice"]},"model":"glm-5.2","effort":"low","cost_usd":0.0,"raw_usage":{"total_tokens":745,"prompt_tokens":488,"completion_tokens":257,"prompt_tokens_details":null},"tokens_in":488,"tokens_out":257,"duration_ms":4750,"temperature":1.0,"reasoning_tokens":63,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-09T20:25:17.572504+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If at larger system sizes the detuned-staggered field develops entanglement-asymmetry crossings (or the staggered field loses them), the claimed separation between spectral chaos and Mpemba crossings would narrow or collapse.","supporting_citations":[],"review_version":1}