{"id":"e99f929a-f1a1-4e62-a791-9126523817ca","arxiv_id":"2509.01931","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For systems with projective symmetry representations, the paper proposes a modified ETH and shows that charged operators with symmetry-supplied charges thermalize to a generalized Gibbs ensemble, not the ordinary Gibbs ensemble.","lead":"This paper extends the Eigenstate Thermalization Hypothesis, which explains how isolated quantum systems reach equilibrium, to systems whose symmetries wind around each other with a phase (projective representations, often tied to 't Hooft anomalies). It shows that certain 'charged' observables keep a memory of the initial state and that their late-time values are captured by a generalized Gibbs ensemble rather than the standard thermal ensemble.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"prETH ansatz (3.16) is not established for Type II operators with nonlocal neutralized forms; Appendix A shows the ansatz can fail, so the GGE claim's domain is under-specified.","rationale":"The reader identified the prETH ansatz, Eq. (3.16), as the load-bearing assumption, and I agree. My stress-test sharpens this: the ansatz is not merely unproved; it is demonstrably false in related settings (Appendix A.2), and the paper's numerical evidence for Type II operators only covers the special case where the neutralized operator is local (Eq. 3.21). In that special case, prETH reduces to standard local ETH, so the central GGE claim is on much firmer ground than the general ansatz suggests. The concrete test directly probes the regime where prETH is most vulnerable—nonlocal neutralized operators—and would settle whether the GGE description extends beyond the 'local neutral times symmetry operator' construction. The reader's CONDITIONAL verdict remains appropriate: the argument is coherent and numerically supported for the stated examples, but the general claim is restricted by this unvalidated assumption. No verdict change is needed.","tokens_in":38694,"tokens_out":20642,"duration_ms":237130,"concrete_test":"In the Sec. 5.2 Z2 gauge theory (Lx odd, Ly even), take a neutral nonlocal operator A with O(1) norm, e.g., the product of two non-contractible Wilson loops W_x W_y, and define a charged operator O_{q1,q2} = A (U1)^{q2}(U2)^{-q1} so that its neutralized form is exactly A. Compute the diagonal matrix elements ⟨⟨Ei||A||Ei⟩⟩ in the prETH basis for increasing Lx, Ly, and test whether they collapse to a smooth function of E/V with fluctuations scaling as e^{-S/2}. If they do not, or if the long-time average of O_{q1,q2} differs from Eq. (3.37) by more than O(V^{-1/2}), then prETH and the GGE prediction fail beyond the local-neutralized Type II class.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on the prETH ansatz, Eq. (3.16): the neutralized matrix element ⟨⟨Ei||O_{q1,q2}(U2)^{q1}(U1)^{-q2}||Ej⟩⟩ is asserted to have a smooth diagonal O^{(q1,q2)}(E/V) and exponentially suppressed off-diagonals. This is the step that converts the exact selection-rule expression (3.15) into the state-independent stationary value (3.37), and then into the GGE equality (4.20). However, the ansatz is not universally valid even within the paper's own framework: Appendix A.2 shows that for Z_{N1}×Z_{N2} with N1≠N2, operators involving the center elements (U1)^n(U2)^n violate the diagonal smoothness of prETH (Eq. A.21). In the N1=N2 numerical tests, the only Type II operators used are of the special form O_{0,0}(U1)^{q2}(U2)^{-q1} (Eq. 3.21), for which the neutralized operator reduces to the local neutral O_{0,0}, so prETH reduces to standard local ETH. No test is reported for a Type II operator whose neutralized operator is genuinely nonlocal. If prETH fails for such operators, Eq. (3.37) and the GGE description of their stationary values fail. The paper thus establishes the GGE claim only for the restricted class of Type II operators with local neutralized counterparts, not for all Type II operators as the general ansatz suggests.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a projective-representation generalization of the Eigenstate Thermalization Hypothesis (prETH) for isolated quantum systems with Abelian Z_{N1} × Z_{N2} symmetries acting projectively. The projective structure forces degeneracies in the energy spectrum, and the authors derive an exact selection-rule expression for the long-time average of charged operators, Eq. (3.15). They then conjecture the prETH ansatz, Eq. (3.16), for the neutralized matrix elements, and classify charged operators as neutral, Type I (vanishing diagonal in the thermodynamic limit), or Type II (nonvanishing diagonal). The main claim is that Type II operators retain memory of the projective charge of the initial state through the factor <ψ_in|(U_1)^{q_2}(U_2)^{-q_1}|ψ_in>, so their stationary values are not described by the standard Gibbs ensemble but by a noncommuting generalized Gibbs ensemble, Eqs. (4.10)-(4.20). The paper supports this with exact diagonalization studies in Z_2 × Z_2 and Z_3 × Z_3 spin chains and in a (2+1)-dimensional Z_2 lattice gauge theory, and discusses anomalous O(V^{-1/2}) finite-size corrections in Section 6. The general Z_{N1} × Z_{N2} case is treated in Appendix A, where the authors themselves exhibit a violation of diagonal prETH for operators involving center elements.","tokens_in":39209,"tokens_out":7090,"duration_ms":71476,"significance":"If the prETH ansatz is accepted, the paper gives a clean and interesting extension of ETH to degenerate spectra induced by projective representations and 't Hooft anomalies, and it demonstrates a concrete mechanism by which the standard Gibbs ensemble fails while a noncommutative GGE succeeds. The exact selection-rule time-average, Eq. (3.15), and the GGE matching calculation are valuable and appear correct. The numerical evidence, while limited to small systems, is consistent with the claimed smooth diagonal structure and with GGE matching for the selected operators. However, the central predictive claim is conditional on a conjectural ansatz whose domain is not fully specified: the paper's own Appendix A shows that the diagonal prETH can fail for certain symmetry-structure operators, and the numerical Type II examples all reduce to local neutralized operators. The significance is therefore real but conditional on a sharper statement of the validity domain.","major_comments":[{"comment":"The central result is only as general as the prETH ansatz, and that ansatz is not universal even within the paper's own framework. For G = Z_{N1} × Z_{N2} with N1 ≠ N2, Eq. (A.21) shows that operators of the form O_{q'1,q'2}(U_1)^n(U_2)^n have diagonal matrix elements equal to a sector-dependent phase times the neutralized diagonal element, so at least one of the two related operators cannot have a smooth diagonal function of E/V. The main-text Type II numerical tests (Figs. 1, 2, 5) are all of the special form (3.21), where the neutralized operator is just the local neutral O_{0,0}; no test exercises a genuinely nonlocal neutralized operator. Consequently Eq. (3.37) and the GGE statement (4.20) are established only for the subclass in which (3.16) is assumed. The abstract and Section 4 state a more general Type II claim. Please state the domain of validity explicitly, and either prove s","section":"§3.2, Eq. (3.16), and Appendix A.2, Eq. (A.21)"},{"comment":"The equality between the long-time average and the GGE is substantially built in by construction. The GGE parameters are fixed by Eq. (4.14) to match <ψ_in|(U_1)^{q_2}(U_2)^{-q_1}|ψ_in> for all (q_1,q_2), and the Type II GGE expectation value in Eq. (4.19) is proportional to exactly this same quantity. Thus Eq. (4.20) follows from the matching conditions once prETH supplies the smooth diagonal function O^{(q1,q2)}. This does not make the derivation wrong, but it means the GGE is not making an independent prediction of the stationary value; the nontrivial content is the prETH ansatz and the nonvanishing of O^{(q1,q2)}. The paper should state this limitation explicitly in Section 4, otherwise readers may overinterpret the GGE agreement as a stronger test than it is.","section":"§4.2, Eqs. (4.14)-(4.20)"},{"comment":"The classification of local charged operators as Type I is a conjecture, supported only by a heuristic argument about the support of the neutralized operator and by a few numerical examples. This distinction is load-bearing: if a local charged operator were actually Type II, the Gibbs-ensemble prediction for that operator would fail, and the Type I/II boundary would move. The current evidence covers only Z_1, X_1 X_2 U_1, σ^x_ℓ, and W_y for particular couplings and system sizes. I recommend either a more systematic numerical study (several local operators at several sizes, with explicit scaling of the diagonal matrix elements) or a more rigorous locality-based argument, so that the classification is not a per-operator numerical observation.","section":"§3.2, Type I conjecture, Eqs. (3.18)-(3.19)"}],"minor_comments":[{"comment":"The matrix element in Eq. (3.33) is written as <⟨E_i|| O_{q1,q2} ||E_i⟩>, but Eq. (3.15) and the surrounding text require the neutralized operator O_{q1,q2}(U_2)^{q1}(U_1)^{-q2} inside the matrix element. Please correct this notation to avoid ambiguity.","section":"§3.2, Eq. (3.33)"},{"comment":"There is an inconsistency in the labeling of Type I and Type II operators. The text defines O_{0,1}^{I}=Z_1 and O_{0,1}^{II}=X_1^† X_2 U_1, but the Fig. 2 caption appears to swap these labels. Please make the notation uniform.","section":"§5.1, Z_3 × Z_3 example and Fig. 2 caption"},{"comment":"The bound in Eq. (A.26) is displayed as an equality O(V^{-1/2}) for all n ≥ 1, but for n ≥ 2 the standard estimate gives O(V^{-1}) under the same assumptions. Since Eq. (A.27) only needs an upper bound, this does not affect the final result, but the displayed equality should be corrected.","section":"Appendix A.2, Eq. (A.26)"},{"comment":"There are several small presentation issues: 'bahaviors' in Section 1.3 should be 'behaviors'; Z_N × Z_N spacing is inconsistent in places; and the notation <⟨E_i|| ... ||E_j⟩> with double angle brackets should be defined once and used consistently.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern largely lands: the paper's own Appendix A narrows the validity of the prETH ansatz, while the main-text numerical tests for Type II operators all use the special form (3.21). The GGE-vs-long-time-average agreement is also partly built into the parameter-matching conditions. These issues are fixable by restricting the claims appropriately, adding tests for genuinely nonlocal neutralized Type II operators, and clarifying the logical status of the GGE 'prediction'. I do not see grounds for rejection, but the abstract and Section 4 currently overstate the generality of the result."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi — bottom line: this paper is worth reading. It formulates ETH for systems with projective Abelian symmetry and derives a concrete consequence: charged operators of \"Type II\" (charge supplied by symmetry operators) do not thermalize to Gibbs, but to a noncommutative GGE. The exact selection rule (3.15) is correct, and the GGE matching in Sec. 4 follows cleanly. The three-way classification of charged operators is new and useful. The numerical tests in spin chains and Z2 gauge theory are consistent, and the gauge theory example with mixed 1-form/0-form symmetry is a nice concrete demonstration.\n\nThe soft spots are real but not fatal. The whole structure rests on the prETH ansatz (3.16), which is a conjecture. That's fine for an ETH paper — ETH itself is a conjecture — but the numerics only test Type II operators of the special form O_{q1,q2}=O_{0,0}(U1)^q2(U2)^-q1, where the neutralized operator reduces to local O_{0,0}. So prETH is being verified in a regime where it reduces to standard local ETH. Appendix A shows the ansatz can fail for genuinely nonlocal neutralized operators in the N1≠N2 case. Thus the paper establishes the GGE claim for a restricted class, not for all Type II as the abstract implies. The GGE fitting also has an element of built-in consistency: chemical potentials are chosen to reproduce the same charges that enter the time average. That is standard GGE logic, but it means the predictive content is exactly \"these charges are the only relevant ones.\"\n\nThere's a small factor-of-2 slip in the anomalous-scaling example (Sec. 6.2): the state's <U2> is 1/2, not 1/4, and the stationary value should be (1/2) O(...) rather than (1/4). Worth fixing but doesn't affect the main argument.\n\nWho gains: people working on ETH in gauge theories and on 't Hooft anomalies in statistical mechanics; also lattice gauge theory numerics. I'd send it to a serious referee. The correct verdict is conditional: the core idea is likely right, but the paper needs to state clearly where prETH is expected to hold, ideally with a test of a Type II operator whose neutralized form is nonlocal.\n\nRecommendation: engage with it; referee it.","headline":"A solid, interesting extension of ETH to projective Abelian symmetries; the GGE result is clean but rests on an ansatz whose domain is narrower than the numerics demonstrate.","tokens_in":39563,"tokens_out":5644,"would_cite":true,"duration_ms":57806,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["05.30.-d"],"model":"deepseek-v4-flash","headline":"When Abelian symmetries act with projective phases instead of commuting, thermalization of charged observables is governed by a generalized Gibbs ensemble that keeps a memory of the initial state's projective charge, not the standard Gibbs","keywords":["eigenstate thermalization hypothesis","projective representation","mixed anomaly","generalized Gibbs ensemble","thermalization","degenerate energy eigenstates","lattice gauge theory","higher-form symmetry"],"falsifier":"Exact diagonalization of larger systems (beyond L = 13 for the spin chains and 3×4 for the gauge theory): if the neutralized diagonal ⟨⟨Ei||O_{q1,q2}(U2)^{q1}(U1)^{-q2}||Ei⟩⟩ for a Type II operator stops being a smooth O(1) function of energy density, or if the long-time average deviates from Eq. (3.37) by more than O(V^{-1/2}), the prETH ansatz fails. Sharper test: prepare the engineered state |ψan⟩ of Eq. (6.23) and measure a Type II stationary value; it should shift by O(V^{-1/2}) relative to the mean-energy prediction — if the shift scales as V^{-1}, the anomalous-scaling mechanism is wron","tokens_in":38633,"feed_emoji":"⚛️","tokens_out":12236,"duration_ms":116481,"temperature":0.7,"pith_summary":"The paper extends the Eigenstate Thermalization Hypothesis to quantum systems whose Abelian symmetry groups act projectively — symmetry operators that commute only up to a U(1) phase, a signature of mixed quantum anomalies. It argues that such projective structures force every energy eigenstate into a degenerate multiplet, and it formulates a projective-representation ETH (prETH) whose matrix-element ansatz applies to the symmetry-neutralized combination of a charged operator with the symmetry generators. The central consequence is a three-way operator classification: neutral and Type I charged operators still thermalize to the standard Gibbs ensemble, whereas Type II charged operators — whose charge is supplied by the symmetry operators themselves — settle at a stationary value that carries the initial state's projective charge. That value is reproduced by a non-commutative generalized Gibbs ensemble, not by the Gibbs ensemble, which selection rules force to zero for every charged operator. If prETH holds, dephasing alone cannot erase the memory of anomalous symmetry charge in highly excited states.","feed_headline":"Projective charges make thermal states remember the initial state","feed_subtitle":"Charged observables settle into a generalized Gibbs ensemble that remembers the initial projective charge.","key_machinery":"The load-bearing object is the projective commutation relation U2U1 = e^{−2πi/N}U1U2, the minimal nontrivial projective phase for Z_N × Z_N corresponding to a mixed anomaly; it forces every energy eigenstate into an N-fold degenerate multiplet. On top of this sits the prETH ansatz (Eq. 3.16): the symmetry-neutralized matrix element ⟨⟨Ei||O_{q1,q2}(U2)^{q1}(U1)^{-q2}||Ej⟩⟩ has the standard ETH shape, with a smooth diagonal function O^{(q1,q2)}(E/V) and exponentially suppressed off-diagonal noise. This ansatz converts the exact selection-rule factorization of the long-time average into the stationary value (3.37). The matching ensemble is the non-commutative GGE (Eq. 4.10), exp(−βH − Σ μ_r Q_r","core_discovery":"Central claim: when Z_N × Z_N symmetry acts projectively, the long-time average of a Type II charged operator equals the initial state's expectation value of the symmetry operator (U1)^{q2}(U2)^{-q1} times a smooth function of energy density, up to O(V^{-1/2}) corrections (Eq. 3.37). The factorization follows exactly from the degeneracy selection rules; the prETH ansatz (Eq. 3.16) — smooth diagonal, exponentially suppressed off-diagonal neutralized matrix elements — makes the stationary value state-independent. Selection rules force the Gibbs ensemble to give zero for every charged operator, so Type II observables cannot thermalize to Gibbs. The paper instead builds a non-commutative general","pith_inferences":["If prETH holds, the diagonal ensemble provides a symmetry-protected quantum memory — the value ⟨ψin|(U1)^{q2}(U2)^{-q1}|ψin⟩ survives thermalization and is readable through any Type II observable, suggesting anomalous systems could store quantum information in highly excited states.","The anomalous O(V^{-1/2}) scaling could serve as an experimental diagnostic: measuring how a Type II operator's stationary value approaches its infinite-volume GGE value reveals the projective-charge content of the initial state without full state tomography.","The prETH logic should transfer to other anomalous group structures — central extensions of larger Abelian groups, or mixed 0-form/higher-form anomalies — wherever a neutralized matrix element can be defined; the paper's appendix already takes a first step for Z_{N1} × Z_{N2} with N1 ≠ N2.","A sharp boundary question is whether every local charged operator is genuinely Type I; if a local counterexample with non-vanishing neutralized diagonal exists, the GGE predictions would need revision, making the Type I conjecture an independently testable claim."],"forward_implications":["Type II charged operators — charge supplied by the symmetry operators themselves — equilibrate to a value carrying ⟨ψin|(U1)^{q2}(U2)^{-q1}|ψin⟩, so the stationary state retains exact memory of the initial projective charge even after dephasing.","The standard Gibbs ensemble is provably wrong for these observables; the non-commutative GGE reproduces them to O(V^{-1/2}), so thermal equilibrium in anomalous systems means generalized-Gibbs equilibrium, not Gibbs.","Neutral and Type I operators still follow the conventional Gibbs prediction, so familiar ETH thermalization survives exactly within the neutralized sector.","Type II observables show anomalous finite-size corrections of order V^{-1/2} rather than the usual V^{-1}; such corrections are exponentially atypical among random initial states but can be engineered by superposing few energy eigenstates, as in the explicit state |ψan⟩.","The predictions reach beyond abstract chains: the Z_2 lattice gauge theory with odd L_x or L_y realizes the projective structure through its 0-form and electric 1-form symmetry operators, placing the effect inside physical gauge theories."],"supporting_citations":[{"why":"Supplies the finite-size ETH matrix-element ansatz that prETH directly generalizes.","marker":"[1]"},{"why":"Establishes the higher-form-symmetry ETH setting and the argument that Gibbs-thermalization breakdown need not stem from a small bath when symmetries are higher-form.","marker":"[34]"},{"why":"Provides the non-Abelian ETH formulation on which prETH is modeled, including the anomalous finite-size scaling phenomenon.","marker":"[36]"},{"why":"Supplies the non-commutative generalized Gibbs ensemble formalism used to construct ρGGE with non-commuting charges.","marker":"[41]"},{"why":"Provides the Z_N × Z_N clock-and-shift operator construction underlying the spin-chain numerical tests.","marker":"[42]"},{"why":"Supplies the lattice gauge theory framework used for the Z_2 gauge-theory example.","marker":"[43]"},{"why":"Precursor treatment of projective-phase effects on ETH, establishing the degeneracy argument that prETH builds on.","marker":"[44]"},{"why":"Supplies the typicality and effective-dimension argument used to show anomalous scaling is exponentially atypical.","marker":"[45]"}],"fun_headline_variants":["Projective symmetries change ETH: thermal states remember initial charge","ETH with anomalies: charged observables settle to generalized Gibbs","Anomalous symmetry breaks standard thermalization, keeps memory","Quantum thermalization remembers initial state under projective charges","Projective ETH: Gibbs ensemble is not enough for charged operators"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is the prETH ansatz, Eq. (3.16): that the symmetry-neutralized matrix element of any charged operator takes the standard ETH form, with a smooth diagonal function and exponentially suppressed off-diagonal entries — a conjecture the paper checks numerically for a few selected operators but does not derive, and the whole generalized-Gibbs picture collapses if arbitrary charged operators violate it.","fun_headline_variants_meta":{"raw":{"variants":["Projective symmetries change ETH: thermal states remember initial charge","ETH with anomalies: charged observables settle to generalized Gibbs","Anomalous symmetry breaks standard thermalization, keeps memory","Quantum thermalization remembers initial state under projective charges","Projective ETH: Gibbs ensemble is not enough for charged operators"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000504,"raw_usage":{"total_tokens":2265,"prompt_tokens":676,"completion_tokens":1589,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":420,"completion_tokens_details":{"reasoning_tokens":1522}},"tokens_in":420,"tokens_out":1589,"duration_ms":12668,"temperature":1.0,"reasoning_tokens":1522,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:04:54.888617+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exact diagonalization of larger systems (beyond L = 13 for the spin chains and 3×4 for the gauge theory): if the neutralized diagonal ⟨⟨Ei||O_{q1,q2}(U2)^{q1}(U1)^{-q2}||Ei⟩⟩ for a Type II operator stops being a smooth O(1) function of energy density, or if the long-time average deviates from Eq. (3.37) by more than O(V^{-1/2}), the prETH ansatz fails. Sharper test: prepare the engineered state |ψan⟩ of Eq. (6.23) and measure a Type II stationary value; it should shift by O(V^{-1/2}) relative to the mean-energy prediction — if the shift scales as V^{-1}, the anomalous-scaling mechanism is wron","supporting_citations":[{"cited_title":"The approach to thermal equilibrium in quantized chaotic systems,","cited_arxiv_id":null,"evidence_quote":"Supplies the finite-size ETH matrix-element ansatz that prETH directly generalizes."},{"cited_title":"An Introduction to Lattice Gauge Theory and Spin Systems,","cited_arxiv_id":null,"evidence_quote":"Supplies the lattice gauge theory framework used for the Z_2 gauge-theory example."},{"cited_title":"Remarks on effects of projective phase on eigenstate thermalization hypothesis","cited_arxiv_id":"2310.11425","evidence_quote":"Precursor treatment of projective-phase effects on ETH, establishing the degeneracy argument that prETH builds on."}],"review_version":1}