{"id":"a84b950d-b6b7-4e95-a2ff-5be5f2ea4791","arxiv_id":"2501.11302","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The paper derives the entanglement entropy of two disjoint intervals in a thermal 2D conformal field theory, identifies its holographic dual as an entanglement wedge cross section in a BTZ black hole, and shows this cross section can cross the horizon in the thermofield double state.","lead":"Two physicists compute how much quantum entanglement exists between two separated pieces of a hot two-dimensional quantum system, using a subtraction trick instead of the usual purification. They match the result to a geodesic in a black hole geometry and show that, in a two-sided thermofield double state, this geodesic can pass through the black hole horizon.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The exactness of Eq. (16) rests on dropping all excited states in the annulus partition function; this is uncontrolled when the conformal width W is small, so the claimed equality S_vN=EWCS is not established.","rationale":"The reader's verdict is CONDITIONAL, and my analysis supports that: the CFT derivation of the central formula has a genuine gap. However, I do not think the weakest point is the Affleck-Ludwig boundary entropy itself. Substituting Eq. (12) into Eq. (3) gives S^(n) = (c/12)(1+1/n)W + log⟨a|0⟩⟨0|b⟩, so the boundary overlap appears as a constant additive term. If it is independent of the annulus moduli (as boundary-state overlaps usually are), subtracting it in Eq. (14) is a consistent normalization, not an unjustified removal. The real problem is the truncation of the k-sum in Eq. (12): the partition function of the annulus contains infinitely many conformal dimensions, and the vacuum term is only the leading contribution for large W. The paper applies the result to all values of the cross ratio ζ, including ζ → 0 where W → 0, and in that regime excited states are not suppressed. If those states contribute, the Rényi entropy would carry additional ζ-dependent, model-dependent terms, and the equality with the exact bulk EWCS would fail. The bulk computation itself is straightforward and the horizon-crossing analysis in Section III appears internally consistent, but the claimed CFT-side universality is not established. A direct check using an exactly solvable CFT (free boson/fermion) would settle whether Eq. (16) is exact or only a large-W approximation. This is why the verdict remains CONDITIONAL rather than ACCEPT: the main claim may be true, but the derivation as written omits a controlled justification for the vacuum truncation.","tokens_in":10660,"tokens_out":14791,"duration_ms":156641,"concrete_test":"Compute the exact Rényi entropy for two disjoint intervals in a thermal free-boson or free-fermion CFT using the standard twist-field four-point function on the torus, and take the n → 1 limit. Compare the resulting S_vN(A:B) with Eq. (16) over a range of ζ, including ζ ≪ 1 (where W ≪ 1). If the exact result differs from (c/6) arccosh(1+2ζ) by any ζ-dependent term, Eq. (25) fails. Alternatively, evaluate the annulus partition functions Z1 and Zn in Eq. (8) and Eq. (11) keeping the full k-sum for a compact boson and check whether the n → 1 limit of S^(n) is boundary-condition independent and equals c/6 W.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central derivation replaces the full annulus partition function Z1 = e^{cW/12} Σ_k ⟨a|k⟩⟨k|b⟩ e^{-2δ_k W} (Eq. (8)) by the k=0 vacuum term in Eq. (12). The suppression of excited states requires W ≫ 1/δ_{k>0}, but the final formula (16) is asserted for all cross ratios ζ ∈ (0,∞). From Eq. (15), ζ → 0 gives R2/R1 → 1 and W → 0, where the vacuum-dominated expansion is not valid. The same issue affects Zn in Eq. (11), whose effective width is W/n, making excited states even more relevant as n → 1. If excited states contribute, S^(n) acquires boundary-condition- and spectrum-dependent terms that do not cancel in the n → 1 limit, and the equality S_vN(A:B) = (c/6) arccosh(1+2ζ) would not follow from the subtraction construction. The Affleck-Ludwig overlap ⟨a|0⟩⟨0|b⟩ flagged by the reader is a constant that can be absorbed by the subtraction in Eq. (14), so it is not the primary obstruction; the uncontrolled truncation of the k-sum is. The paper's statement that 'the contribution of excited states is irrelevant' (after Eq. (12)) is an assertion, not a derivation, and it is load-bearing because the bulk EWCS (Eq. (23)) is exact for all ζ.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript extends a previously introduced 'subtraction approach' to mixed-state entanglement in a thermal CFT2 on a cylinder. After mapping the cylinder to the plane and removing the separating intervals C and D by cutting out two discs, the remaining subsystem A∪B is treated as living on an annulus. The authors compute the von Neumann entropy of this pure-state configuration from the annulus partition function, drop all excited-state contributions and the Affleck-Ludwig boundary entropy, and obtain S_vN(A:B) = (c/6) log(1 + 2ζ + 2√(ζ(ζ+1))) with ζ defined in Eq. (17). They compare this with the entanglement wedge cross section computed from geodesics in the planar BTZ black hole and claim equality in Eq. (25). They also study a two-sided thermofield double configuration and derive a horizon-crossing condition κ=1 in Eq. (34).","tokens_in":11025,"tokens_out":13689,"duration_ms":134935,"significance":"If established, the result would provide a simple finite-temperature check of the subtraction prescription and an unambiguous CFT-side characterization of the holographic entanglement wedge cross section, extending the zero-temperature results of Refs. [16,17]. The bulk calculation in Appendix A is explicit and self-contained, and the comparison involves no fitted parameters: the CFT expression and the geodesic computation are independent and agree in analytic form. The horizon-crossing feature in the thermofield double state, if correct, is an interesting qualitative prediction. However, the significance is currently limited by the lack of control over the CFT approximation, and the paper should not be accepted before that gap is addressed.","major_comments":[{"comment":"The replacement of the full annulus partition functions (8) and (11) by their k=0 vacuum terms is the load-bearing step of the whole CFT computation, but it is only asserted. Suppression of excited states requires W = log(R2/R1) to be large compared with 1/δ_{k>0}; however, Eq. (15) implies W→0 as ζ→0, a regime explicitly covered by the final formula (16). In that regime the k-sum is not vacuum-dominated. The same issue affects the replicated partition function (11), whose effective width is W/n, so the n→1 limit does not restore vacuum dominance. No large-central-charge or sparse-spectrum limit is stated that would justify vacuum dominance in a holographic CFT. Since the bulk EWCS in Eq. (23) is exact for all ζ, the claimed equality (25) is not demonstrated. The authors should either prove that excited-state contributions cancel after the subtraction or state the precise domain of validity of Eq. (16) and explain how the comparison to the exact bulk result is meaningful.","section":"II.A, Eq. (12)"},{"comment":"The Affleck-Ludwig boundary entropy log(⟨a|0⟩⟨0|b⟩) is removed with the statement that it is 'totally irrelevant' to the entanglement between A and B. This is a definitional choice, not a consequence of the replica computation: in Eq. (13) this term is independent of n, so it survives in the von Neumann limit and would affect the value of S_vN(A:B) unless cancelled by an additional prescription. If the overlap depends on the intervals that define the boundary conditions, or on the annulus modulus, Eq. (14) and therefore Eq. (16) contain an unremoved model-dependent contribution. The authors should derive the cancellation or state explicitly the normalization convention under which the boundary entropy is discarded.","section":"II.A, Eqs. (13)-(14)"},{"comment":"The two-sided configuration is one of the advertised new results, but the modified cross ratio (30) is introduced without derivation. It is not immediate how Eqs. (16) and (19) 'can be directly read off' for points living on two different asymptotic boundaries of the thermofield double state, and the coordinate transformation (22) as presented covers one exterior patch of the BTZ geometry. The horizon-crossing analysis in Eqs. (33)-(34) therefore needs a self-contained calculation in the extended BTZ spacetime, including an explanation of how the geodesic endpoints are continued across the horizon and why that geodesic remains the minimal entanglement wedge cross section.","section":"III, Eq. (30)"}],"minor_comments":[{"comment":"There are several typographical errors: 'statenth Rényi entropy', 'ER = EP Rhas', and the spacing in 'ρAA∗BB ∗' should be corrected.","section":"I"},{"comment":"The double limit x2-x1→ϵ and x4-x3→ϵ is written loosely, and the constant inside the logarithm after the limit is cutoff-dependent; please specify the regularization and comment on the factor 2π/β relative to the standard thermal interval entropy.","section":"Eq. (20)"},{"comment":"The term 'phase transition' for the crossing at ζ=1 is potentially misleading: the paper shows that the green geodesics belong to different bipartitions, so the crossing is a comparison of two different quantities rather than a phase transition in a single mixed-state measure; this should be stated more carefully.","section":"II.C"},{"comment":"The symbol w is used both for the coordinate on the original w-plane and for the boundary coordinate in Eq. (32); please introduce separate notations to avoid confusion in the horizon-crossing formulas.","section":"Eqs. (31)-(34)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a sequel to the authors' own prior works [16,17], and the subtraction prescription itself is not independently scrutinized here; the referee report focuses on internal consistency. If the excited-state truncation cannot be justified, or the final formula must be restricted to a valid regime, the paper would need substantial reworking. The editors may wish to ensure that the revision addresses the CFT derivation rather than only cosmetic issues."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper extends the same group's subtraction approach from zero temperature to thermal CFT2, and it does one genuinely new thing: it shows that in the TFD state, with intervals on opposite boundaries, the EWCS can cross the BTZ horizon when kappa < 1. The CFT side gives S = (c/6) log(1+2zeta+2 sqrt(zeta(zeta+1))) with zeta the thermal cross ratio, and the bulk geodesic computation matches that expression. The phase-transition discussion at zeta=1 and the identification of which geodesic corresponds to which boundary pair are clean and useful.\n\nThe problem is the CFT derivation. The annulus partition function (8) is truncated to the vacuum in Eq. (12) with the statement that excited states are 'irrelevant.' No derivation is given. The Affleck-Ludwig boundary term is a constant and can indeed be subtracted, so the reader's worry about that is minor. But the truncation of the k-sum is not minor. The suppression factor e^{-2 delta_k W} only works for large conformal width W, and W = log(1+2zeta+2 sqrt(zeta(zeta+1))) goes to zero as zeta->0. In that regime the vacuum block is not dominant, and the equality S=EW is not established. The paper asserts the formula for all zeta in (0,infty). This is load-bearing because the bulk EWCS is exact for all zeta.\n\nIf the authors can show that excited states cancel in the n->1 limit after the subtraction, or restrict the claim to a region where W is large, the paper would be solid. As it stands, the central equality is a conjecture supported by the bulk match, not a derivation.\n\nWho should read this? People working on mixed-state entanglement and EWCS in AdS/CFT, especially those interested in the TFD setup. It is worth a serious referee, because the horizon-crossing result and the phase-transition discussion are interesting and likely correct. But the referee should press on the vacuum-dominance step.\n\nMy recommendation: send to peer review, but require the authors to state the regime of validity or prove the cancellation.","headline":"Thermal extension of the subtraction approach matches EWCS, but the CFT derivation rests on an unproved vacuum-dominance truncation.","tokens_in":11502,"tokens_out":4166,"would_cite":false,"duration_ms":43052,"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 a bipartite mixed state in a thermal CFT$_2$, the subtraction approach gives $S_{vN}(A:B)=\\frac{c}{6}\\log\\left(1+2\\zeta+2\\sqrt{\\zeta(\\zeta+1)}\\right)$, and the paper proves this equals the entanglement wedge cross section in planar…","keywords":["mixed state entanglement entropy","thermal CFT2","subtraction approach","entanglement wedge cross section","BTZ black hole","thermofield double state","replica trick","boundary entropy"],"falsifier":"On an annulus with two slits representing $A$ and $B$, compute the exact Rényi entropy $S^{(n)}$ for a free compactified boson with general conformal boundary states, keeping the full boundary-state sum; if the $n\\to1$ limit after subtracting the fixed boundary entropy is not $\\frac{c}{6}\\log\\left(1+2\\zeta+2\\sqrt{\\zeta(\\zeta+1)}\\right)$ for all annulus moduli, the central equality is false.","tokens_in":10469,"feed_emoji":"🕳️","tokens_out":15719,"duration_ms":132175,"temperature":0.7,"pith_summary":"The paper aims to establish the bipartite mixed-state entanglement entropy of a thermal two-dimensional conformal field theory from the subtraction approach, and to match it unambiguously to a bulk quantity in the dual BTZ black hole. The core result is that for two disjoint intervals $A$ and $B$, the entropy is $S_{vN}(A:B)=\\frac{c}{6}\\log\\left(1+2\\zeta+2\\sqrt{\\zeta(\\zeta+1)}\\right)$, where $\\zeta$ is a cross ratio of the four endpoints; the complementary pair $C:D$ obeys the same formula with $\\zeta\\to1/\\zeta$. In planar BTZ this is exactly the length of the entanglement wedge cross section, so the paper claims the subtraction prescription is a no-optimization CFT definition of that holographic quantity. It then studies the thermofield double state and shows that, when a parameter $\\kappa$ defined by the interval endpoints is less than one, the dual surface for $C:D$ crosses the black hole horizon. If correct, this gives a concrete, calculationally tractable handle on mixed-state entanglement in holography, with a sharp geometric prediction that can be checked either in CFT or in the bulk.","feed_headline":"Thermal CFT2 mixed-state entropy equals a BTZ geodesic length","feed_subtitle":"A subtraction calculation without optimization pins the holographic dual and finds it crossing the horizon in the thermofield double.","key_machinery":"The mechanism is the subtraction map from the thermal cylinder to an annulus. One maps the cylinder to the plane with $w=e^{2\\pi z/\\beta}$, then removes the separating intervals $C$ and $D$ by cutting out discs whose boundaries carry conformal boundary states $|a\\rangle$ and $|b\\rangle$; the remaining surface is a pure-state annulus with slits for $A$ and $B$. The annulus partition function is $Z_1=e^{cW/12}\\sum_k \\langle a|k\\rangle\\langle k|b\\rangle e^{-2\\delta_k W}$, and the $n$-fold replica is again an annulus with conformal width $W/n$. Keeping only the vacuum term and subtracting the boundary-entropy term $\\log(\\langle a|0\\rangle\\langle 0|b\\rangle)$ gives $S_{vN}(A:B)=cW/6$, which is rewritten in the conformally invariant cross ratio $\\zeta$ through $e^W=1+2\\zeta+2\\sqrt{\\zeta(\\zeta+1)}$. On the gravity side, the same cross ratio enters the geodesic length in planar BTZ, and the horizon-crossing analysis is done by mapping BTZ to flat-slice coordinates and locating where an endpoint coordinate vanishes.","core_discovery":"The central claim is that for a thermal CFT$_2$ at inverse temperature $\\beta$, the subtraction approach gives the static bipartite mixed-state entanglement entropy $S_{vN}(A:B)=\\frac{c}{6}\\log\\left(1+2\\zeta+2\\sqrt{\\zeta(\\zeta+1)}\\right)$, where $\\zeta=\\frac{\\sinh(\\pi(x_2-x_1)/\\beta)\\sinh(\\pi(x_4-x_3)/\\beta)}{\\sinh(\\pi(x_3-x_2)/\\beta)\\sinh(\\pi(x_4-x_1)/\\beta)}$ for intervals $A=(x_1,x_2)$ and $B=(x_3,x_4)$. The same expression with $\\zeta\\to1/\\zeta$ gives the entropy of the complementary pair $C=(x_2,x_3)$ and $D=(-\\infty,x_1)\\cup(x_4,\\infty)$. The paper shows, using the planar BTZ metric and the standard central-charge relation, that the entanglement wedge cross section for the same cross ratio is exactly this CFT answer, so $S_{vN}(A:B)=EW(A:B)$. For the thermofield double state, with the modified cross ratio of Eq. (30), the dual $EW(C:D)$ crosses the horizon when $\\kappa<1$ and $EW(A:B)$ does so when $\\bar{\\kappa}<1$, with $\\kappa$ and $\\bar{\\kappa}$ given explicitly in Eqs. (34) and (36).","pith_inferences":["If the subtraction result is universal, it provides an optimization-free CFT definition of the entanglement wedge cross section that could extend to multipartite mixed states and to other holographic settings.","The horizon-crossing transition at $\\kappa=1$ and $\\bar{\\kappa}=1$ is a sharp prediction that could be compared with reflected entropy or negativity in the same thermofield double configuration; those quantities should show a qualitative change at the same parameter value if the geometric picture is complete.","Because the final formula depends only on the cross ratio, a natural test is to evaluate the exact annulus Rényi entropy in free-boson or minimal-model boundary conditions; if the residual boundary contribution varies with the annulus moduli, the equality holds only for special boundary conditions."],"forward_implications":["As a consequence of the claimed equality, the entanglement wedge cross section is the holographic dual of the subtraction-approach entropy in thermal states, extending the zero-temperature verification.","The subtraction approach fixes the phase transition between $S_{vN}(A:B)$ and $S_{vN}(C:D)$ at $\\zeta=1$ without an optimization procedure, so each boundary pair is matched to a definite bulk geodesic.","In the thermofield double state, the dual surface for $C:D$ crosses the BTZ horizon when $\\kappa<1$, so mixed-state entanglement in the TFD state is not always confined to one exterior.","The covariant extension follows by replacing $x_i$ with $z_i=x_i+\\tau_i$, so the same formula applies to time-dependent intervals.","The pure-state single-interval entropy is recovered by taking the regulator limit in $S_{vN}(C:D)$, yielding the standard thermal result $\\frac{c}{3}\\log\\left(\\frac{\\beta}{\\pi\\epsilon}\\sinh\\frac{\\pi\\ell}{\\beta}\\right)$."],"supporting_citations":[{"why":"Introduces the subtraction approach for mixed states in CFT at zero temperature and confirms the equality with the entanglement wedge cross section, the method and target the present paper extends to thermal states.","marker":"[16, 17]"},{"why":"Hypothesizes the entanglement wedge cross section as the holographic dual of mixed-state entanglement, the equality the paper verifies in planar BTZ.","marker":"[11]"},{"why":"Supplies the annulus partition function as a sum over boundary states that is the starting point of the replica calculation.","marker":"[18–20]"},{"why":"Identifies the boundary entropy term that the paper subtracts as irrelevant to the entanglement between $A$ and $B$.","marker":"[21]"},{"why":"Provides the central-charge to Newton-constant relation used to convert the BTZ geodesic length into the entropy formula.","marker":"[22]"},{"why":"Studies a two-sided thermofield double configuration that the paper compares with its own setup and extends with the horizon-crossing analysis.","marker":"[9]"}],"fun_headline_variants":["Thermal CFT2 entropy: subtraction matches BTZ exactly","No optimization: thermal CFT2 entropy equals BTZ geodesic","Horizon-crossing entropy for thermofield double in CFT2","Exact mixed-state entropy: thermal CFT2 without extremization","Thermal CFT2: subtraction gives exact holographic entropy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the boundary-entropy term left by removing the separating intervals is completely irrelevant to the entanglement between $A$ and $B$, so that dropping it after Eq. (13) leaves a universal result.","fun_headline_variants_meta":{"raw":{"variants":["Thermal CFT2 entropy: subtraction matches BTZ exactly","No optimization: thermal CFT2 entropy equals BTZ geodesic","Horizon-crossing entropy for thermofield double in CFT2","Exact mixed-state entropy: thermal CFT2 without extremization","Thermal CFT2: subtraction gives exact holographic entropy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001189,"raw_usage":{"total_tokens":4887,"prompt_tokens":904,"completion_tokens":3983,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":520,"completion_tokens_details":{"reasoning_tokens":3893}},"tokens_in":520,"tokens_out":3983,"duration_ms":26319,"temperature":1.0,"reasoning_tokens":3893,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T18:26:14.644300+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On an annulus with two slits representing $A$ and $B$, compute the exact Rényi entropy $S^{(n)}$ for a free compactified boson with general conformal boundary states, keeping the full boundary-state sum; if the $n\\to1$ limit after subtracting the fixed boundary entropy is not $\\frac{c}{6}\\log\\left(1+2\\zeta+2\\sqrt{\\zeta(\\zeta+1)}\\right)$ for all annulus moduli, the central equality is false.","supporting_citations":[],"review_version":1}