{"id":"f883ad36-e971-445e-bb47-79fc6e4ac8a5","arxiv_id":"2607.18050","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For d ≤ min{S1,S2}, every nonseparable Werner state with parameter Φ ∈ [−(d−1)/min{S1,S2}, 0) satisfies all Bell inequalities under any S1×S2-setting scenario with generalized measurements; for d > min{S1,S2}, every nonseparable Werner state does.","lead":"This paper derives a new analytic rule for when a nonseparable Werner state (a standard example of an entangled-but-possibly-local quantum state) remains Bell-local when the two parties are limited to S1 and S2 measurement settings. The rule extends the range of the state parameter beyond the classic Werner and Barrett bounds for many setting counts, and re-derives a 2003 optimization result of Terhal, Doherty, and Schwab with explicit operators.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central bound depends on unverified minimal eigenvalue in Eq. (27); direct numerical check needed","rationale":"The reader's primary weakest assumption was Theorem 1/Proposition 2 imported from prior work. On inspection, the specific application here only needs the standard extension argument for positive source operators, which is straightforward and correct. The more concrete load-bearing input is the minimal eigenvalue of the flip-operator sum in Eq. (27), because the numerical content of the central bound (19) is exactly Φ ≥ −(d−1)/Smin, and this number is τ = λ_min/S2. The paper cites [22,23] rather than proving it, and those references were unavailable to the reader. A cheap numerical check for small dimensions would settle it. The algebraic error in Eq. (26) is a real typo but immaterial to the bound. Therefore I do not see a reason to alter the reader's conditional verdict; the same concerns remain, with the eigenvalue verification being the most useful condition to impose.","tokens_in":11023,"tokens_out":20283,"duration_ms":195183,"concrete_test":"Compute the spectrum of J_{d,m} = Σ_{j=2}^m F_{1j} for d ∈ {2,3,4} and m ∈ {3,4,5,6} by exact diagonalization (or symbolic algebra) and confirm min eigenvalue = −min(d−1, m−1). If any deviation appears, re-check the positivity interval in Theorem 2(a) and the bound (19). Additionally, re-derive Eq. (26) including the 1/(1−τ) factor.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 2(a)'s interval (19) is derived from the positivity threshold γ ≥ 0, which requires Φ ≥ τ = λ_min(J_{d,S2+1})/S2 with λ_min = −min(d−1,S2). This eigenvalue fact is imported from [22,23] without proof. If the true λ_min were less negative, the claimed locality interval would overreach; if more negative, the constructed operator would not be positive on the full claimed interval. The rest of the proof is internally consistent once Eq. (26) is corrected by the missing factor 1/(1−τ) (as the reader notes; this typo does not affect the bound). The extension→LHV step is also sound for the positive source operators, since one can build a joint distribution on A and the S2 B-copies and take λ = (b_1,...,b_{S2}). Thus the single most load-bearing external input is the minimal eigenvalue of J.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies S1×S2-setting Bell locality of nonseparable Werner states. Using the author's earlier framework of source operators and tensor positivity, it constructs explicit positive 1×S2 and S1×1 source operators for W_{d,Φ} by taking a projection onto the minimal eigenspace of J = Σ_j F_{1j}. The main result, Theorem 2, has three parts: (a) for 2≤d≤min{S1,S2}, W_{d,Φ} is S1×S2-setting Bell local whenever Φ∈[−(d−1)/min{S1,S2},0), with additional locality for larger numbers of settings; (b) for min{S1,S2}<d≤max{S1,S2}, every nonseparable Werner state is S1×S2-setting Bell local; (c) for d>max{S1,S2}, the same holds with extensions on both sides. The paper also gives an explicit operator proof of the Terhal–Doherty–Schwab optimization result for d>min{S1,S2}.","tokens_in":11002,"tokens_out":20273,"duration_ms":171692,"significance":"If the theorem is correct, it provides an analytic, parameter-free sufficient condition for Bell locality of nonseparable Werner states in restricted-setting scenarios, going beyond the Werner and Barrett thresholds for many S1,S2, and it recovers the TDS d>min{S1,S2} result in explicit operator form. The construction is concrete and the partial-trace computations are directly checkable. The main advertised interval (19) appears defensible; however, the theorem as stated contains an overclaim in its 'moreover' parts, and one key spectral input is imported without proof. These issues are local rather than fatal.","major_comments":[{"comment":"The proof states that 'for 2≤d≤Smin, a nonseparable Werner state has positive 1×S2-setting and S1×1-setting source operators if Φ∈[−(d−1)/Smin,0)'. This intersection is computed incorrectly. Positivity of the 1×S2 operator requires Φ≥−(d−1)/S2, and positivity of the S1×1 operator requires Φ≥−(d−1)/S1. The intersection of these two intervals is Φ≥−(d−1)/max{S1,S2}, not Φ≥−(d−1)/min{S1,S2}. For example, with d=2, S1=2, S2=100, condition (19) allows Φ=−0.3, at which the 1×100 source operator is not positive. Consequently the 'moreover' claims — 'L1′×S2-setting Bell local for all L1′>S1' (when S1=Smin) and the symmetric claim — are not established by the given argument. The basic S1×S2 locality claim remains supported because the source operator on the smaller side is positive under (19), but the theorem statement and proof need correction: either the 'moreover' parts should be stated under","section":"III, Theorem 2(a) and proof after Eq. (28)"},{"comment":"The bound (19) is exactly the assertion that the minimal eigenvalue of J_{d,S2+1} is −min{d−1,S2}. This is the single numerical input that sets the claimed locality interval, but it is imported from Refs. [22,23] with no proof or statement of the spectrum. Given that the new bound is no stronger than this fact, the manuscript should include a short derivation (e.g., from the representation theory of the symmetric group) or at least a precise statement of the eigenvalue formula with a fully self-contained reference. As currently written, this is a load-bearing omitted proof.","section":"III, Eq. (27)"}],"minor_comments":[{"comment":"The expression for γ is missing the positive factor 1/(1−τ_{P_{d,S2+1}}). Inserting (24)–(25) into the symmetric-component equation gives γ=(Φ−τ_{P})/[(1−τ_{P}) · binom(d+S2,S2+1)]. The error does not change the positivity threshold γ≥0, since 1−τ_{P}>0 for the chosen P=P_{λmin}, but the formula as printed is incorrect.","section":"III, Eq. (26)"},{"comment":"The text says 'the Barrett's locality bound γ_gm^B(2)=1/8 in (5)'; Eq. (5) is Werner's bound, while Barrett's bound appears in Eq. (6). Please correct the cross-reference.","section":"IV, Conclusion"},{"comment":"Minor typographical issues: 'noseparable' in the Introduction; 'Terhal et. el.' in the abstract and references; 'semi-programming' should presumably be 'semidefinite programming'. These do not affect the mathematics.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The central construction and the main S1×S2 locality interval are credible, and the explicit positive source operators are a useful contribution. The main issue is the overclaim in Theorem 2(a)'s 'moreover' parts, which follows from an incorrect interval-intersection step; this is fixable without changing the core construction. I would also encourage the author to make the paper more self-contained regarding the minimal-eigenvalue fact and the implication from positive source operators to Bell locality, since both are imported from prior work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper likely proves a genuinely new sufficient condition for S1×S2-setting Bell locality of Werner states with d ≤ Smin, and gives a clean operator proof of the Terhal–Doherty–Schwab result for d > Smin. The main construction checks out—I re-derived the source operator coefficients and the only printed mistake I found is a missing factor 1/(1−τ_P) in Eq. (26), which does not affect the bound. The comparisons to Werner and Barrett are honest, and the numerical examples are correct. So the core claim is credible.\n\nWhere I'd push back: the proof leans on Theorem 1 and Proposition 2 imported from the author's earlier work. For the positive source operators actually built here, the standard extension-to-LHV argument works, so this dependence is benign in this application. The real load-bearing external input is the minimal eigenvalue of J_{d,m} being −min{d−1,m−1}, cited to [22,23] without proof. If that fact is wrong, the claimed interval (19) could overreach. I did not have those refs at hand; a direct check of the eigenvalue for small d and S2 would settle it. The relation to asymmetric-extendibility results in Jakab et al. (2022) should also be clarified—does (19) follow from or improve on those thresholds?\n\nThe paper is short, clearly written, and the author is explicit about what is new (d≥3, and d=2 with Smin≥4) and what reproduces TDS. That transparency counts. I would send it to a referee, mainly to verify the eigenvalue citation and to confirm the imported theorems' applicability. It is not a desk-reject.","headline":"New Werner-state locality bound is real, but the proof rests on an unverified eigenvalue fact; worth a careful referee.","tokens_in":11761,"tokens_out":2157,"would_cite":true,"duration_ms":20682,"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":"For nonseparable Werner states, a new analytic bound decides when they satisfy all Bell inequalities under S1×S2-setting scenarios.","keywords":["Werner state","Bell locality","Bell inequalities","source operator","tensor positivity","measurement settings","local hidden variable model","nonlocality"],"falsifier":"Take d=2, S1=S2=7, and Φ=−0.14, which lies in the claimed local interval [−1/7,0). If any 7×7-setting Bell inequality is violated by the Werner state W_{2,Φ}, Theorem 2(a) is false. Equivalently, one can numerically check whether the constructed operator T = γP^{(+)}+ξP_{λmin} with these parameters is positive; if it is not, the proof's pivot fails.","tokens_in":10674,"feed_emoji":"⚛️","tokens_out":12362,"duration_ms":89424,"temperature":0.7,"pith_summary":"The paper asks when a nonseparable Werner state—the standard mixed entangled state—actually displays Bell nonlocality when only limited measurement settings are available at each site. It proves a new sufficient condition: if the state's dimension d is no larger than the smaller of the two setting counts S1, S2, and the mixing parameter Φ lies in [−(d−1)/min{S1,S2}, 0), then the state satisfies every Bell inequality for all such S1×S2-setting scenarios, with outcomes of any spectral type. This condition is stronger than previously known locality bounds for a wide range of settings (for example, for d=2 it beats the old bounds whenever min{S1,S2}≤7). The paper also shows that when d exceeds min{S1,S2}, every nonseparable Werner state is automatically Bell local under those settings, re-deriving an earlier optimization result in explicit operator form. The practical consequence is that a Werner state's nonlocality can be hidden simply by choosing the number of measurement settings appropriately.","feed_headline":"Nonseparable Werner states turn local once both sides use d settings","feed_subtitle":"The threshold (d−1)/min(S1,S2) determines when a Werner state hides its entanglement from Bell tests.","key_machinery":"The construction uses an operator T = γP^{(+)}_{d,S2+1} + ξP_{λmin} on H_d ⊗ H_d^{⊗S2}, where P^{(+)}_{d,S2+1} projects onto the fully symmetric subspace of S2+1 copies and P_{λmin} projects onto the eigenspace of J = Σ_{j=2}^{S2+1}F_{1j} (sum of flip operators between the first and j-th copies) with its minimal eigenvalue. The coefficients γ and ξ are chosen so that the partial trace of T equals the Werner state. The crucial spectral fact is that the minimal eigenvalue of J is −min{d−1, S2}, so the operator is positive exactly when Φ lies above −min{d−1,S2}/S2. This eigenvalue computation converts a positivity question into the explicit interval (19) of the theorem.","core_discovery":"The central claim, Theorem 2, has three parts. For 2≤d≤Smin, a nonseparable Werner state W_{d,Φ} is S1×S2-setting Bell local whenever Φ∈[−(d−1)/Smin, 0); more strongly, it is also local for any number of settings at the other site beyond S2 (or at this site beyond S1). For Smin<d≤Smax, every nonseparable Werner state is S1×S2-setting Bell local, and the same holds for d>Smax. The proof is constructive: it exhibits explicit positive 'source operators'—extensions of the state to multiple copies—whose partial trace returns W_{d,Φ}. Positivity of these operators is what grants Bell locality.","pith_inferences":["The same flip-operator eigenvalue mechanism might be applied to other symmetric families of states (such as isotropic states) to produce setting-dependent locality thresholds; the present paper does not explore this.","The threshold (d−1)/Smin is only sufficient; whether it is tight—that is, whether Werner states with Φ just below it violate some S1×S2-setting Bell inequality—remains open.","The result implies a general caution: increasing the number of measurement settings in a Bell test does not always make nonlocality easier to detect; for Werner states, more settings can actually force locality.","In a multipartite setting, a similar construction with a star of flip operators could yield locality bounds for N-partite Werner-like states, with the bound depending on the graph spectrum."],"forward_implications":["In any S1×S2-scenario with min{S1,S2}≥d, a Werner state with Φ∈[−(d−1)/min{S1,S2},0) admits a local hidden-variable description: it cannot be used to violate any Bell inequality.","For d=2, the new locality bound beats the previous best bound for all scenarios with min{S1,S2}≤7; for d=3, for min{S1,S2}≤32.","When d>min{S1,S2}, no nonseparable Werner state can show Bell nonlocality at all under those limited scenarios, regardless of Φ.","The explicit positive source operator constructed for d>Smin provides a concrete, state-level witness that the earlier optimization result holds with a fully explicit extension.","Protocols that certify entanglement by Bell-violating Werner states must restrict to fewer than min{S1,S2} settings per site; at or above that number the state is provably local."],"fun_headline_variants":["Werner states turn local at the min(S1,S2) threshold","Bell locality bound for nonseparable Werner states","Nonseparable Werner states pass Bell tests under d settings","New limit: Werner state nonlocality vanishes at min(S1,S2)","S1×S2 Bell locality secured for Werner states below d"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The proof borrows a previously proven theorem stating that a positive 'source operator' (an extension of the state to extra copies) implies Bell locality under all S1×S2 settings; the paper constructs such positive operators but does not re-prove that implication.","fun_headline_variants_meta":{"raw":{"variants":["Werner states turn local at the min(S1,S2) threshold","Bell locality bound for nonseparable Werner states","Nonseparable Werner states pass Bell tests under d settings","New limit: Werner state nonlocality vanishes at min(S1,S2)","S1×S2 Bell locality secured for Werner states below d"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000158,"raw_usage":{"total_tokens":1106,"prompt_tokens":831,"completion_tokens":275,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":187}},"tokens_in":575,"tokens_out":275,"duration_ms":2931,"temperature":1.0,"reasoning_tokens":187,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T16:22:37.175475+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take d=2, S1=S2=7, and Φ=−0.14, which lies in the claimed local interval [−1/7,0). If any 7×7-setting Bell inequality is violated by the Werner state W_{2,Φ}, Theorem 2(a) is false. Equivalently, one can numerically check whether the constructed operator T = γP^{(+)}+ξP_{λmin} with these parameters is positive; if it is not, the proof's pivot fails.","supporting_citations":[],"review_version":1}