{"id":"c768477e-0b09-49b9-8bcc-deddd9100421","arxiv_id":"2511.07212","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Interacting N-state chiral clock chains have exact matrix-product-state eigenstates (the true ground states in gapped regions) whenever the Hamiltonian's Laurent polynomial is a perfect square, and these states form a dense skeleton.","lead":"This paper constructs exact matrix-product-state wavefunctions for a family of interacting spin chains — the N-state chiral clock models — extending a construction that previously worked only for free-fermion systems. It delivers exact ground and excited states in gapped regions, a dense approximation scheme for gapped phases, and a closed-form disorder parameter: new analytic tools for a notoriously hard integrable-model family.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Representation-lift of Eq. (61) from the N=2 Majorana computation to all N is unproved; a d≥2 N=3 check would settle whether Result 1 survives.","rationale":"I read the paper as a serious analytic extension of the MPS-skeleton program. The d=1 pivot algebra, the disorder-parameter transfer-matrix calculation, and the explicit N=2 reduction are all credible and I found no internal inconsistency in those checks. However, the central N>2 claim — that the N=2 Majorana computation fixes a representation-independent formula for the dressed Hamiltonian — rests on an unproved faithfulness assertion. The paper's own wording in Section 5.1 ('provided we do not have any representation-specific cancellations') explicitly flags the gap, and the subsequent range-counting argument does not close it: it addresses relations involving long-range generators, not arbitrary L-independent relations among the short-range operators that actually appear in Eq. (61). This is exactly the reader's weakest assumption, and I agree that it is the most load-bearing point: if the lift fails, Results 1–3 for N>2 do not follow. The proposed N=3, d=2 exact-diagonalization check directly tests the lifted identity and the resulting eigenstate, so it would settle whether the concern lands. I therefore recommend CONDITIONAL acceptance: the paper's analytic structure is sound and likely correct, but the N>2 theorem should be contingent on this concrete confirmation, after which ACCEPT would be appropriate. I did not raise a separate objection about Result 4's 'lowest-energy' wording, because it is secondary and does not affect the central skeleton/density claims.","tokens_in":28404,"tokens_out":17012,"duration_ms":186829,"concrete_test":"Use exact diagonalization for N=3, L=6 (and repeat for N=4, L=8 if feasible), d=2, with s=(1,0.3,0.2), so g(z)=1+0.3z+0.2z^2 and f(z)=g(z)^2. Construct A_0 and A_1 via Eq. (9); generate A_2,A_3,A_4 from the pivot relations Eq. (12), and obtain G_α for α=−1..2 from [A_0,A_α] together with G_α=−G_{−α}. Compute b_2=s_2/s_0=0.2 and b_1 via Algorithm 1, then β_k=2 arctanh(b_k), and evaluate H̃_A = e^{β_2 A_2} e^{β_1 A_1} H_A e^{-β_1 A_1} e^{-β_2 A_2}. Compute the operator norm of H̃_A − Σ_{α=−1}^{2} rα(Aα+Gα) using Algorithm 2's rα, normalized by ||H̃_A||, after subtracting any overall identity component. If this residual is not below ~1e−10, the Eq. (61) lift fails. As a cross-check, in the same run verify H_A e^{−β_2 A_2}e^{−β_1 A_1}|ψ_0> is proportional to e^{−β_2 A_2}e^{−β_1 A_1}|ψ_0>.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing step is Section 5.1's assertion that the transformed Hamiltonian computed in the N=2 Majorana representation, H̃_A = Σ_{α=1−d}^{d} rα(Aα + Gα) (Eq. 61), holds in every Onsager representation. The proof computes the coefficients rα in the N=2 representation, Eq. (60), and argues that no representation-specific cancellations can occur because any linear dependence among the short-range operators {A_l, G_m : |l|,|m|≤d} would have to involve order-L generators. That inference is not justified. Finite-dimensional representations of the infinite-dimensional Onsager algebra generally have relations — the generator set cannot be faithful on a finite Hilbert space — and the paper neither exhibits a basis of the relevant universal-algebra subspace nor proves injectivity of the N=2 representation on it. The L≫d range argument only excludes relations involving generators with index ~L; it does not rule out a fixed, L-independent relation among the short-range set. If such a relation exists in the Majorana representation but is absent in the N=3/4 clock representations, the coefficients read off from Eq. (60) are representation-specific and Eq. (61) fails for N>2; then Result 1 — and hence Results 2 and 3 — collapse. The d=1 case is not a test of this: Eq. (24) follows algebraically from the pivot relation for every N, so the critical case is d≥2, where the range-counting argument is doing the work.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies N-state chiral clock chains whose Hamiltonians are linear combinations of Onsager algebra generators A_m. For Laurent polynomials f(z)=±z^p g(z)^2, it constructs states |φ⟩=M^(d)⋯M^(1)|ψ±_p⟩ with M^(k)=exp(∓2 arctanh(b_k)A_{p+k}) and claims they are exact eigenstates (Result 1), exact ground states when H_A lies in the gapped set S (Result 2), that such 'skeleton' Hamiltonians are dense in S (Result 3), and that analogous exact MPS eigenstates exist for low-lying excitations (Result 4). It also derives ⟨µ_k⟩=√(1−a²) for the disorder parameter in the d=1 family. The proof strategy is to conjugate H_A using the Onsager algebra, compute the coefficients in the N=2 free-fermion representation, and lift the resulting operator identity to all N.","tokens_in":28569,"tokens_out":12207,"duration_ms":109851,"significance":"If valid, these results are a substantial advance: they carry the MPS-skeleton program beyond free-fermion systems into genuinely interacting chiral clock chains, provide exact MPS eigenstates (not only ground states), give a density statement with an explicit approximation scheme, and produce a closed-form disorder parameter. The paper includes reproducible algorithmic definitions (Schur-Cohn coefficients and the r_α routine) and clearly separates what is representation-independent from what is clock-model-specific. The main caveat is the unproved representation-lift of the central operator identity.","major_comments":[{"comment":"The representation-lift is the load-bearing step: r_α are computed in the N=2 Majorana representation, and the paper argues they remain valid for all N by ruling out representation-specific cancellations with a range argument. That argument only excludes relations involving order-L generators; it does not rule out L-independent relations among the short-range set {A_l,G_m: |l|,|m|≤d}. Finite-dimensional Onsager representations do satisfy linear relations (see §4), and the d=1 case is not a test because Eq. (24) follows algebraically. Please prove the needed linear independence (e.g., via the sl(2) decomposition) or verify Eq. (61) numerically for d=2, N=3. Without this, Results 1–3 remain conditional.","section":"§5.1, Eq. (61)"}],"minor_comments":[{"comment":"The parenthetical 'G_k = G_{−k}' is a sign typo. Eq. (51) gives G_{−k} = −G_k, and the d=1 derivation of Eq. (24) uses this antisymmetry.","section":"§5.1, after Eq. (60)"},{"comment":"The notation |φ^(P)_{E_0+1/N, p=0}⟩ is defined only in Eq. (17); consider defining it before Eq. (16) to avoid forward-reference confusion.","section":"§2.1, Eq. (16)"},{"comment":"The displayed circulant matrices T and B are difficult to read because of the ellipses and missing entries. A symbolic definition with F_1,F_2 is available and is clearer; the matrix display can be simplified or removed.","section":"Appendix B"},{"comment":"When |b_k|>1, the path connecting H_A(s) to the fixed-point Hamiltonian A_l is asserted but not spelled out. A short explicit description of the path and why it stays in S would strengthen Result 2.","section":"§5.2.2"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely publishable if the representation-lift in §5.1 is fixed. The reader's accept verdict is more optimistic than my own: the central proof is conditional at present, and the missing piece is checkable. I would encourage the editor to request the d=2, N=3 verification or a representation-theoretic proof of linear independence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this deserves a serious referee. It moves the MPS-skeleton idea from the N=2 free-fermion setting to interacting chiral clock chains, constructing exact MPS eigenstates for f(z)=±z^p g(z)^2 and showing these form a dense skeleton in gapped regions. Results 1–3 are new for N>2, and Result 4 (MPS excited states) is new even at N=2. The disorder-parameter formula for general even N is a concrete payoff.\n\nThe paper is genuinely good in the ways that matter. The central construction is explicit and checkable: the d=1 pivot computation goes through (modulo a sign typo in the parenthetical “G_k=G_{-k}” in Sec. 5.1; Eq. (51) gives G_{-k}=-G_k), and the N=4 disorder-parameter transfer-matrix identity checks out. The authors are also scrupulous about scope: Result 2 is conditional on being in the gapped region S, the |b_k|=1 points are flagged, the disorder formula is proved only for d=1, and the Outlook explicitly says the proof is “not fully independent” of Ref [19]. The citation practice is honest rather than self-promotional.\n\nThe soft spot is exactly the one the stress-test flags. Section 5.1 lifts Eq. (61) from the N=2 Majorana computation to all N by arguing that no representation-specific cancellations can occur: any linear dependence among short-range generators would have to involve order-L generators, which is impossible for L≫d,p. That inference is not justified. Finite-dimensional representations of the infinite-dimensional Onsager algebra can have L-independent relations among small-index generators; the range argument does not exclude them. If such a relation exists in the N=2 Majorana representation but not in the N=3 or N=4 clock representations, the coefficients r_alpha read off from Eq. (60) would be representation-specific and Eq. (61) would fail for N>2, taking Results 1–3 down with it. Nothing in the paper suggests that this actually happens, and the d=1 case is consistent, but d=1 is not a test: Eq. (24) follows algebraically for every N. A d=2, N=3 exact-diagonalization overlap check on a small chain would settle it quickly.\n\nThis is a real gap in the proof, not a manufactured concern, and it is structurally separate from the MPS ansatz itself. I would still accept the paper for peer review with the expectation of revision, and I would cite it for the excited states and the disorder-parameter result. The central claim is plausible and likely correct, but the representation-lift step needs either a better proof or a numerical check before I would call it airtight.","headline":"A serious, worthwhile generalization of the MPS-skeleton program to interacting N>2 clock chains, but the proof of the key representation-lifting step is not fully closed and needs either an argument or a small numerical check.","tokens_in":29283,"tokens_out":1886,"would_cite":true,"duration_ms":22347,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that interacting N-state chiral clock chains have a dense family of Hamiltonians with exact matrix-product ground states, extending the free-fermion MPS skeleton to non-free models.","keywords":["MPS skeleton","Onsager algebra","chiral clock models","exact matrix-product ground states","integrable quantum chains","disorder parameter","Schur-Cohn coefficients","dense skeleton"],"falsifier":"Numerically diagonalize a small chiral-clock chain with N=3 (or N=4) for a skeleton Hamiltonian with degree d=2 polynomial g (e.g., g(z)=1+a z + c z^2 with all |b_k|≠1) and check whether the explicitly constructed MPS |φ⟩ satisfies H_A|φ⟩ = E|φ⟩ up to machine precision. If it fails for a generic such case, the representation-lifting assumption breaks.","tokens_in":28043,"feed_emoji":"⚛️","tokens_out":6356,"duration_ms":50411,"temperature":0.7,"pith_summary":"The paper asks whether the 'MPS skeleton'—a dense network of Hamiltonians whose ground states are exactly matrix-product states—exists beyond free-fermion models. It answers yes for N-state Onsager-integrable chiral clock chains: whenever the Hamiltonian's Laurent polynomial f(z) factors as ±z^p g(z)^2, the model has an exact eigenstate of the form exp(−β_d A_{p+d})…exp(−β_1 A_{p+1})|ψ±_p⟩, which is an MPS. In the gapped regions surrounding the fixed-point generators A_k, this state is the ground state, and the skeleton is dense there, giving an analytic MPS approximation to any gapped ground state. The paper also constructs exact MPS excited states for even p and, as an application, computes the disorder parameter ⟨µ_k⟩ = √(1−a²) for the family A_0 + 2a A_1 + a² A_2 for all even N. Much of the argument uses only the Onsager algebra and is representation-independent, going beyond the chiral-clock setting.","feed_headline":"Interacting clock chains get exact MPS ground states","feed_subtitle":"Beyond free fermions: exact MPS ground states and a closed disorder parameter in interacting chains.","key_machinery":"The Onsager algebra with generators {A_m, G_m} and the pivot relation e^{β A_m} A_l e^{-β A_m} = cosh²(β/2) A_l + sinh(β) G_{m-l} − sinh²(β/2) A_{2m-l}. Conjugating the skeleton Hamiltonian by a product of such exponentials reduces it (in the free-fermion representation and via an argument that the form is representation-independent) to a combination Σ r_α (A_α + G_α) that leaves the fixed-point ground state |ψ_0⟩ invariant. The coefficients b_k from the Schur-Cohn algorithm control the parameters β_k; the resulting state is an MPS because each exponential is a local MPO acting on a product state.","core_discovery":"For Hamiltonians H_A = Σ t_m A_m built from the Onsager algebra generators A_m in the N-state chiral clock representation, if the Laurent polynomial f(z) = Σ t_m z^m has the factorized form f(z) = ±z^p g(z)^2 with g a polynomial, then H_A has an exact eigenstate |φ⟩ = exp(−β_d A_{p+d})...exp(−β_1 A_{p+1})|ψ±_p⟩, where β_k = 2 arctanh(b_k) and the b_k are Schur-Cohn coefficients of g. This state is an exact MPS. When H_A lies in a gapped region around a fixed-point generator A_k, |φ⟩ is the ground state. The skeleton is dense in those gapped regions, so any gapped ground state can be approached by such MPS. For even p, analogous MPS eigenstates give low-lying excited states. As an application","pith_inferences":["We infer that the skeleton may give a new route to correlation functions beyond the superintegrable line: taking limits along skeleton paths (as the paper suggests for determinantal formulas) could yield closed expressions for order/disorder parameters in more general Onsager-integrable chains.","We infer that the exact excited-state MPS can serve as a diagnostic for phase boundaries: the gap in a gapped region must close when one of these excited states becomes degenerate with the ground state, giving concrete variational bounds on the boundary location.","We infer that, by analogy with the N=2 case where the |b_k|≠1 restriction was removed in later work, the corresponding singular cases for N>2 might also admit MPS ground states through limiting arguments, extending the skeleton to the full measure-zero set.","We infer that the MPS forms constructed here could be used as high-quality initial states for numerical algorithms and as a testbed for exploring frustration-free properties in interacting chains."],"forward_implications":["For every N, the chiral-clock Hamiltonians whose Laurent polynomial is a perfect square (up to a monomial and sign) have exact MPS eigenstates, so these interacting integrable chains share an analytic structure previously known only for the N=2 free-fermion class.","The MPS skeleton is dense in the gapped regions around the Onsager generators: any gapped ground state in those regions can be approximated arbitrarily closely in energy density by an explicit MPS state, providing an analytic complement to numerical tensor-network methods.","In the d=1 family H_A = A_0 + 2a A_1 + a^2 A_2, the disorder parameter equals √(1−a²) for every even N, extending the known N=2 result to interacting chains.","For even p, the lowest excited states at zero and finite momentum are also exact MPS eigenstates, a feature that goes beyond the ground-state-only analysis of the free-fermion skeleton.","Because the core steps rely on the Onsager algebra, the same construction applies in principle to any finite-dimensional representation with appropriate properties, not only the chiral clock models."],"fun_headline_variants":["Exact MPS ground states in interacting clock chains","Dense MPS skeleton emerges in interacting clock chains","Beyond free fermions: exact MPS states in clock chains","Onsager algebra yields exact MPS states in spin chains","MPS skeleton dense in gapped regions of interacting chains"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The claim that the transformed Hamiltonian's coefficient form, computed in the N=2 Majorana representation, survives in every other representation of the Onsager algebra—that is, the short-range operators {A_l, G_m} have no additional linear relations in the chiral-clock representations.","fun_headline_variants_meta":{"raw":{"variants":["Exact MPS ground states in interacting clock chains","Dense MPS skeleton emerges in interacting clock chains","Beyond free fermions: exact MPS states in clock chains","Onsager algebra yields exact MPS states in spin chains","MPS skeleton dense in gapped regions of interacting chains"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000991,"raw_usage":{"total_tokens":4093,"prompt_tokens":853,"completion_tokens":3240,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":597,"completion_tokens_details":{"reasoning_tokens":3173}},"tokens_in":597,"tokens_out":3240,"duration_ms":18473,"temperature":1.0,"reasoning_tokens":3173,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T23:08:22.766492+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Numerically diagonalize a small chiral-clock chain with N=3 (or N=4) for a skeleton Hamiltonian with degree d=2 polynomial g (e.g., g(z)=1+a z + c z^2 with all |b_k|≠1) and check whether the explicitly constructed MPS |φ⟩ satisfies H_A|φ⟩ = E|φ⟩ up to machine precision. If it fails for a generic such case, the representation-lifting assumption breaks.","supporting_citations":[],"review_version":1}