{"id":"75281d45-b044-4c2b-aba4-c8f4baedfce3","arxiv_id":"2411.09427","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a stack of NS5-branes, the entanglement entropy of radiation grows only like 1/sqrt(N) times time, so early evaporation is suppressed at large N, while island saddles restore it later.","lead":"The paper studies very large stacks of NS5-branes in string theory and reports that in this limit the growth of entanglement entropy, and therefore early evaporation, is strongly suppressed. At much later times, quantum island corrections take over and evaporation proceeds normally, according to the paper.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Late-time island claim is unsubstantiated: the area term A(∂I)/(4G_N) in Eq. (3.37) is never computed and only the matter entropy is extremized, so the conclusion that evaporation becomes unsuppressed at late times has no support.","rationale":"The reader's verdict of CONDITIONAL is appropriate. My concern elevates one of the reader's listed assumptions to the primary blocking issue: the late-time island contribution is not just a numerical detail, it is the second half of the abstract's claim. The early-time result (3.26) is a straightforward consequence of the LST temperature T=1/(2π l_s √N): the Kruskal coefficient c_LST/r0 equals the surface gravity, so the linear growth rate (c/3)κ is indeed 1/√N suppressed. That part of the paper is internally consistent. However, the 'at much later times evaporation proceeds as usual' clause rests entirely on the assertion S(χ)≈A/(4G_N) in (3.37). The generalized entropy (3.18) requires an extremum of A/(4G_N)+S_matter, but only S_matter is extremized. The area term depends on the island position through the metric (2.7), and without it we cannot tell whether the saddle survives the large-N limit. In particular, if A/(4G_N) contains a positive power of N in the denominator, the late-time saddle would also be suppressed and the paper's central dichotomy would fail. The concrete test above settles this. Thus the verdict stays CONDITIONAL pending this computation.","tokens_in":9770,"tokens_out":10601,"duration_ms":98502,"concrete_test":"Compute A(∂I)/(4G_N) for the boundary at u1=r0(1+1/(2 y2^2)) using the full 10D metric (2.7), including the S^3 volume and transverse directions, then extremize the full generalized entropy (3.18) over y1 and y2. If the full extremum differs from (3.36) or the area term is O(N^{-p}) with p>0, the late-time island saddle (3.37) does not dominate at large N; if the extremum is reproduced and the area unsuppressed, the claim survives.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section 3.2.1 defines the island entropy as S(χ)=A(∂I)/(4G_N)+S^a_matter (Eq. 3.30). The extremization yielding (3.36) is performed only on the matter term; the area A(∂I)/(4G_N) is not computed or varied. After substituting the matter extremum, the paper asserts S(χ)≈A(∂I)/(4G_N) (Eq. 3.37) without verifying that this point extremizes the full generalized entropy (3.18). Since the candidate boundary y1≈1+1/(2 y2^2) lies outside the horizon, A includes the S^3 volume factor of the metric (2.7), whose ratio to G_N is never evaluated. Therefore the central large-N claim that the new saddles are no longer suppressed, stated in the abstract and in Section 5, is unsupported by the calculation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript studies the time-dependent entanglement entropy of a stack of N near-extremal NS5-branes in the decoupled (LST) limit, using the island prescription of Almheiri–Mahajan–Maldacena. The author writes the ten-dimensional supergravity metric (2.7) in Kruskal coordinates (2.11)–(2.13), fixes the constant c_LST = sqrt(m_s^2 r_0^2 / N) by requiring the conformal factor to be regular at the horizon, and feeds it into the universal 2d CFT interval-entropy formula (3.21). The single-interval ('Hawking') entropy then grows at rate dS/dt_y = (c/3)(1/sqrt(N))(1/l_s), Eq. (3.26), which is interpreted as a large-N suppression of the early evaporation process. The island section considers two disjoint intervals, approximates the matter entropy (3.33) in two regimes, claims an extremum at t_y1 = t_y2, epsilon = 1/(2 y_2^2), and asserts S(chi) ≈ A(dI)/(4G_N) (3.37), from which it concludes that the late-time saddles are no longer suppressed at large N. Section 4 attempts to show that the suppression is not a coordinate artifact, and Section 6 repeats the computation for an SQCD-like background. The early-time derivation is internally consistent given its stated 2d s-wave assumptions; the late-time conclusion is not supported by the calculation as written, for the reasons detailed below.","tokens_in":9939,"tokens_out":48485,"duration_ms":431846,"significance":"If the early-time result (3.26) is correct, it is a novel and striking large-N effect: the NS5/LST system would decouple from its radiation bath at large N, echoing the zero-width meson argument from large-N QCD cited in the introduction. The central 1/sqrt(N) factor is derived rather than fitted: c_LST is fixed by regularity of the Kruskal conformal factor (2.15), the standard smoothness condition in two-sided black-hole analyses, and no free parameter is tuned to enforce the suppression. The paper is also commendably explicit about its own circumscribed setting: the s-wave reduction, the 2d CFT formula, and the footnote after (3.18) noting that the Page curve obtained is that of a non-gravitating theory. The principal limitation is that the late-time half of the abstract's claim rests on an uncomputed area term; if that gap is closed and the normalization and central-charge questions raised below are answered, this would be a worthwhile contribution to the holographic information-paradox literature. As it stands, the paper presents an interesting candidate mechanism whose second half is not yet demonstrated.","major_comments":[{"comment":"The late-time claim that the island saddles are 'no longer suppressed at large-N' is not established. The variation leading to (3.36) is performed on the matter term S^a_matter alone; the area term A(dI)/(4G_N) in (3.30) is never evaluated for the LST metric (2.7) nor included in the extremization. A stationary point of the sum (3.18) is not in general a stationary point of one summand, so (3.37) does not follow. Moreover, at the candidate values (3.36) the matter term is not small: at the extremum S^a_matter ≈ c log y_2 + O(1), which grows with the bath size y_2, so dropping it in (3.37) requires an explicit comparison with A/(4G_N). The needed area involves the S^3 of radius sqrt(N)/m_s in (2.7), whose volume is proportional to N^{3/2}; the ratio of this area to G_N is never computed, and without it nothing can be said about the N-dependence of (3.37). Finally, the paper's own caveats conflict with (3.37): the footnote after (3.18) states that the Page curve obtained is 'that of a non-gravitating theory ... and not that of the black hole', and the final remark of Section 5 says that '(3.18) does not take gravitational effects into account'. Both statements describe a computation without the gravitational area term, i.e., precisely the computation actually performed.","section":"Sec. 3.2.1, Eqs. (3.30)–(3.37); abstract; Sec. 5"},{"comment":"The normalization of the entropy and the N-dependence of its prefactor are load-bearing but never stated. Eq. (3.19) is introduced as the 2d CFT matter entropy 'per unit area' of the metric (2.13), yet (3.26) is used as the entropy S(chi) with no multiplication by, or reduction of, the transverse volume. The LST metric (2.7) has an S^3 of radius sqrt(N)/m_s and volume proportional to N^{3/2}; under the literal 'per unit area' reading, the total entropy growth rate would acquire a factor N^{3/2} and would grow, not vanish, at large N. If instead c in (3.19) is meant to be the central charge of the reduced s-wave theory, its value and N-dependence must be specified: the suppression in (3.26) is only O(1/sqrt(N)), so a central charge scaling as N^alpha with alpha > 0 would weaken or destroy the effect. Please state the normalization that converts (3.19) into S(chi) and justify the N-independence of the prefactor, specifying whether c is O(1).","section":"Sec. 3.1, Eqs. (3.19), (3.26)"},{"comment":"The extremization as printed cannot be reproduced from the displayed expression. Using Omega(y) = b r_0 y from (2.15), so that d/d(epsilon) log Omega(1+epsilon) ≈ 1 for small epsilon, stationarity of (3.35) with respect to epsilon gives y_2 sqrt(2 epsilon) = 2(1 + epsilon) ≈ 2, i.e., epsilon ≈ 2/y_2^2, a factor 4 larger than the claimed 1/(2 y_2^2). Reproducing the paper's value would require the coefficient of the log Omega term in (3.35) to be 2c/3 rather than c/3, or an additional epsilon-dependence that is not displayed. Since epsilon fixes the position of the island boundary y_1 = 1 + epsilon, and hence the location at which A(dI) would be evaluated, this discrepancy must be resolved; the derivation of (3.35) and (3.36) should be given in full.","section":"Sec. 3.2.1, Eqs. (3.35)–(3.36)"},{"comment":"The argument intended to show that the suppression is not a coordinate artifact is too terse to be a demonstration. The claims that 'the ratio time/space must be invariant under scaling' and that the scale-independence of f_1 forces x_1 and u to scale in the same way, 'inconsistent with the scaling of A_LST', are not supported by any explicit computation of how the Kruskal time t_y appearing in (3.26) is related to a canonically normalized asymptotic time under the proposed rescalings. Furthermore, the abstract's statement that the effect is 'due to the non-invariance of the metric under coordinate parameterisation' is not a well-defined explanation in a diffeomorphism-invariant theory, since the entanglement entropy is a physical quantity and must be invariant. The paper should either present a complete coordinate-invariance check (showing that no rescaling of x_1, u, and t_y together can remove the factor c_LST(N) from the growth rate) or reformulate the claim as a property of a specific parametrization.","section":"Sec. 4, Eqs. (4.39)–(4.42)"}],"minor_comments":[{"comment":"The caption contains a typo: 'Haking radiation' should read 'Hawking radiation'.","section":"Fig. 1"},{"comment":"The string coupling is written as 'g alpha' in 'vanishing string coupling, g_alpha -> 0'; the subscript should be g_s.","section":"Sec. 1"},{"comment":"The heading 'Entaglement entropy for two-disjoint intervals' should read 'Entanglement', and in the introduction 'this new configurations will be complex solutions' should agree in number ('these new configurations').","section":"Sec. 3.2 heading; Sec. 1"},{"comment":"The sentence 'If we assume t_a ≠ t_b this inverts the inequality in (3.34)' uses undefined symbols t_a and t_b; presumably t_{y1} and t_{y2} are meant.","section":"Sec. 3.2.2"},{"comment":"The time scales stated in (3.29) and in Section 5 ('O(t) ~ O(sqrt(m_s^2 r_0^2 / N))') are dimensionally inconsistent as printed: both N m_s^2 r_0^2 and sqrt(m_s^2 r_0^2 / N) are dimensionless, while t_y is a length. Please state the units (e.g., string length) and the per-volume normalization of the entropy used in the comparison with (2.8).","section":"Eq. (3.29); Sec. 5"},{"comment":"Ref. [14] is cited as 'Unpublished, 2023' and underpins the Kruskal-coordinate technology (2.11)–(2.16) used throughout the paper; a published source would be preferable for this load-bearing step.","section":"References"},{"comment":"The constant a is introduced to make arguments dimensionless but its value is never stated; the appearance of F_i(a r) in (2.13) while F_i is defined in (2.12) as a function of x, and the value a = 1/r_0 implied by the time dependence in (3.21), should be clarified explicitly.","section":"Sec. 2.1, Eqs. (2.11)–(2.13)"}],"recommendation":"major_revision","confidential_remarks":"To the editor: I agree with the reader's conditional assessment, and specifically with the stress-test concern: the island/area-term gap in Section 3.2.1 is real and is the main obstacle to the paper's advertised late-time claim. In my view the gap is fixable within the manuscript's scope (compute A(dI)/(4G_N) for the metric (2.7), extremize the full generalized entropy, and compare the result with (3.26)), so I recommend major revision rather than rejection. Two further points: (i) the central Kruskal-coordinate step relies on the author's own unpublished notes [14], which is a weakness in a load-bearing part of the derivation; (ii) if the normalization and central-charge issue raised in Major Comment 2 survives scrutiny, the early-time suppression claim itself may be overturned, so the revision should be checked carefully."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The one thing worth knowing: the early-time entanglement entropy growth for the NS5/LST system does pick up a 1/sqrt(N) suppression, and that part of the paper is easy to verify. The late-time claim that an island saddle removes the suppression is not backed by the calculation; the area term is never computed or extremized.\n\nWhat is actually new: the observation that the Kruskal smoothness constant for the LST metric, c_LST = 1/b with b^2 = N/(m_s^2 r_0^2), turns the standard single-interval entropy into S ~ (c/3)(1/sqrt(N)) t_y/l_s. That is a legitimate extension of the existing island literature, and the paper correctly notes it recovers [26] only in the limit Omega = 1, c = 1. The derivation in Section 3.1 is algebraically clear, and the large-N regime 1 << N << m_s^2 r_0^2 is justified by supergravity constraints. The paper is honest about some limitations, e.g., the footnote that the Page curve obtained is that of a non-gravitating theory satisfying the split property.\n\nThe main problem is Section 3.2. The generalized entropy is S(chi) = A(dI)/(4G_N) + S^a_matter. The extremization in (3.35)-(3.36) is done only on the matter term; the area term A(dI) is never computed for the LST metric, and the full expression is never extremized. Substituting the matter extremum, the paper simply asserts S(chi) ~ A/(4G_N) (3.37). That does not establish the claimed saddle, and it leaves open the possibility that the island contribution is suppressed or absent at large N. So the abstract's late-time statement is unsupported. Also, the 'non-invariance' framing in Section 4 is confusing: any coordinate redefinition that keeps the metric a solution should not change physical quantities; the argument there reads more like a scaling exercise than an invariance statement.\n\nWho this is for: readers working on NS5/LST thermodynamics or island prescriptions in non-AdS backgrounds. It deserves a serious referee, but the referee should ask for a complete treatment of the island area and extremization before the late-time claim can be accepted. The early-time result is worth engaging with on its own.","headline":"A clean early-time large-N suppression result in NS5/LST entanglement entropy, but the late-time island saddle that is claimed to restore evaporation is asserted rather than computed.","tokens_in":10503,"tokens_out":2313,"would_cite":false,"duration_ms":21251,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["04.70.Dy","11.25.-w"],"model":"deepseek-v4-flash","headline":"At early times a high-energy stack of NS5-branes suppresses Hawking evaporation in the large-$N$ limit, with entanglement entropy growing only as $1/\\sqrt{N}$; at much later times island saddles lift the suppression and evaporation…","keywords":["NS5-branes","Little String Theory","entanglement entropy","large-N limit","island formula","Hagedorn temperature","Kruskal coordinates","black hole evaporation"],"falsifier":"Compute the entanglement entropy of the interval in the full ten-dimensional metric (2.7) directly, for example by finding the holographic entanglement surface anchored at the interval endpoints in the numerical background, and compare its early-time slope with $(c/3)(1/\\sqrt{N})(t_y/l_s)$; a slope that does not vanish as $N \\to \\infty$ would falsify the claimed suppression.","tokens_in":9539,"feed_emoji":"🕳️","tokens_out":10993,"duration_ms":101248,"temperature":0.7,"pith_summary":"This paper argues that a black hole made from a large stack of NS5-branes at high energy evaporates anomalously slowly at early times, and that the slowdown is a real large-$N$ effect rather than a coordinate choice. The central quantity is the entanglement entropy between the black hole and its radiation bath: at early times it grows as $S \\sim \\frac{c}{3}\\frac{1}{\\sqrt{N}}\\frac{t_y}{l_s}$, so for $N \\gg 1$ the growth rate tends to zero and the system decouples from the surrounding radiation. The paper's answer to its motivating question is that the suppression is the large-$N$ limit itself, not the Hagedorn temperature, that makes the radiation effectively white noise. At much later times, non-trivial saddles of the gravitational path integral (island configurations) develop an unsuppressed entropy set by $A(\\partial I)/(4G_N)$, and evaporation proceeds in the usual unitary way. The mechanism is the parametric dependence of the Kruskal coordinate constant $c_i$ on $N$; because the metric is not scale-invariant, no coordinate redefinition can remove the suppression, and the same pattern appears in a wrapped-five-brane dual of $\\mathcal{N}=1$ SQCD-like theories.","feed_headline":"Large-N NS5 stacks suppress early Hawking evaporation","feed_subtitle":"Entanglement entropy grows as 1/sqrt(N), so the black hole decouples from its bath until island saddles take over.","key_machinery":"The central object is the constant $c_i$ in the Kruskal transformation $U = -e^{c_i(F_i(x)-ax_1)}$, $V = e^{c_i(F_i(x)+ax_1)}$, chosen so that the conformal factor $\\Omega^2$ in $ds^2 = -\\Omega^2 dU dV + \\cdots$ stays finite at the horizon. For NS5/LST, $c_{\\mathrm{LST}} = 1/b$ with $b^2 = N/(m_s^2 r_0^2)$, so $c_{\\mathrm{LST}} \\sim 1/\\sqrt{N}$ in the large-$N$ window. Because the universal 2d CFT entropy (3.21) contains $c_i t/r_0$ inside a logarithm, this $N$-dependence becomes a $1/\\sqrt{N}$ slope for the early-time entropy. The island contribution is handled with the generalized entropy formula (3.18), whose matter part is the two-interval entanglement entropy (3.31) evaluated near the horizon and near the boundary. The machinery also includes the coordinate-scaling check of Section 4: the functions $f_1$ and $A_{\\mathrm{LST}}$ in (2.7) transform differently under a rescaling, so the suppression cannot be scaled away.","core_discovery":"In the two-sided NS5-brane geometry dual to Little String Theory, the coefficient $c_i$ that enters the Kruskal transformation is not a harmless constant: for the LST background it is $c_{\\mathrm{LST}} = 1/b = \\sqrt{m_s^2 r_0^2/N}$, and the single-interval entanglement entropy inherits this factor in its time dependence, $S_{\\mathrm{matter}} \\sim \\frac{c}{3} c_{\\mathrm{LST}} \\frac{t_y}{r_0} \\sim \\frac{c}{3}\\frac{1}{\\sqrt{N}}\\frac{t_y}{l_s}$. Under the validity bounds $1 \\ll N \\ll m_s^2 r_0^2$, this suppresses the early-time entanglement growth in the large-$N$ limit, so the black hole effectively stops interacting with its radiation bath. The two-interval island computation then shows that at late times the extremum of the generalized entropy is controlled by the area term $A(\\partial I)/(4G_N)$, independent of the $1/\\sqrt{N}$ suppression, so the island saddles dominate and evaporation is no longer suppressed. The paper also verifies that a coordinate rescaling that absorbs $N$ into the metric fails because the time and radial functions in (2.7) scale differently, establishing that the suppression is not an artifact of coordinatisation.","pith_inferences":["The mechanism suggests a search criterion: any gravitational background whose Kruskal conformal factor carries a parameter $c_i(N)$ that tends to zero in some large-$N$ limit will exhibit frozen early-time entanglement growth; scanning known brane and wrapped-brane solutions for such $c_i(N)$ could turn up new examples beyond the two shown here.","If the 2d interval formula genuinely captures the entropy, the suppression is essentially a statement about the relative normalization of time and energy in the holographic dual: the Hagedorn temperature is $N$-independent while the effective energy gap is $1/\\sqrt{N}$, which could be tested by a lattice or matrix-model computation of the spectral form factor.","A direct ten-dimensional holographic entanglement-entropy calculation, anchoring a Ryu-Takayanagi surface at the interval endpoints, would either confirm the $1/\\sqrt{N}$ slope or expose the 2d reduction as the fragile step; this is a concrete next computation implied by the paper's logic."],"forward_implications":["For $1 \\ll N \\ll m_s^2 r_0^2$, the early-time entanglement entropy grows as $(c/3)(1/\\sqrt{N})(t_y/l_s)$; the NS5/LST black hole therefore radiates into its bath at a rate that vanishes as $N \\to \\infty$.","The Page time is estimated as $t_y \\gtrsim N m_s^2 r_0^2$, so the regime of suppressed evaporation can be made arbitrarily long by increasing $N$.","Well before the Page time, the island configuration with $\\epsilon = 1/(2 y_2^2)$ and $t_{y_1}=t_{y_2}$ gives $S(\\chi) \\approx A(\\partial I)/(4G_N)$, an unsuppressed, $N$-independent entropy that restores unitary evaporation.","A rescaling of coordinates cannot remove the $1/\\sqrt{N}$ suppression, because the time and radial pieces of the LST metric carry the $N$-dependence in different ways; the effect is therefore a property of the background, not of the coordinate chart.","The same $1/\\sqrt{N}$ suppression appears in a wrapped-five-brane background dual to $\\mathcal{N}=1$ SQCD-like theories, indicating the phenomenon is not special to the NS5/LST geometry."],"supporting_citations":[{"why":"Establishes the holographic duality between Little String Theory and a two-dimensional black hole, the framework on which the entropy calculation sits.","marker":"[1]"},{"why":"Supplies the island/generalized-entropy extremization formula (3.18) used for the late-time entropy.","marker":"[5]"},{"why":"Provides the NS5/LST holographic duality and the decoupling limit that produces the metric (2.7).","marker":"[10]"},{"why":"Gives the Kruskal-coordinate transformation and conformal factor used to extract the $c_i(N)$ dependence.","marker":"[14]"},{"why":"Provides the universal 2d CFT single-interval entanglement entropy formula used in (3.19).","marker":"[17]"},{"why":"Provides numerical checks supporting the use of the 2d entropy formula in higher-dimensional spacetimes, which the paper relies on to apply (3.19) to the full metric.","marker":"[19]"},{"why":"The linear-dilaton island result that the paper's two-interval expression reduces to when $\\Omega=1$ and $c=1$, serving as the comparison point for the LST calculation.","marker":"[26]"}],"fun_headline_variants":["NS5 stacks: early evaporation fades at large N","Large-N decouples black hole from its bath","1/sqrt(N) stalls black hole evaporation early","Island saddles override large-N evaporation delay","Entanglement entropy reveals large-N evaporation stall"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the universal two-dimensional CFT interval formula, applied to the $(x_1,u)$ slice of the ten-dimensional NS5 metric with the $S^3$ treated as a spectator, gives the full matter entanglement entropy, and that the late-time island configuration extremizes the generalized entropy including the area term $A(\\partial I)/(4G_N)$.","fun_headline_variants_meta":{"raw":{"variants":["NS5 stacks: early evaporation fades at large N","Large-N decouples black hole from its bath","1/sqrt(N) stalls black hole evaporation early","Island saddles override large-N evaporation delay","Entanglement entropy reveals large-N evaporation stall"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001035,"raw_usage":{"total_tokens":4334,"prompt_tokens":897,"completion_tokens":3437,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":513,"completion_tokens_details":{"reasoning_tokens":3363}},"tokens_in":513,"tokens_out":3437,"duration_ms":24842,"temperature":1.0,"reasoning_tokens":3363,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:40:58.830431+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the entanglement entropy of the interval in the full ten-dimensional metric (2.7) directly, for example by finding the holographic entanglement surface anchored at the interval endpoints in the numerical background, and compare its early-time slope with $(c/3)(1/\\sqrt{N})(t_y/l_s)$; a slope that does not vanish as $N \\to \\infty$ would falsify the claimed suppression.","supporting_citations":[{"cited_title":"Callan, Jr., Jeffrey A","cited_arxiv_id":null,"evidence_quote":"Establishes the holographic duality between Little String Theory and a two-dimensional black hole, the framework on which the entropy calculation sits."},{"cited_title":"Islands outside the horizon","cited_arxiv_id":null,"evidence_quote":"Supplies the island/generalized-entropy extremization formula (3.18) used for the late-time entropy."},{"cited_title":"Linear dilatons, NS five-branes and holography.JHEP, 10:004, 1998","cited_arxiv_id":null,"evidence_quote":"Provides the NS5/LST holographic duality and the decoupling limit that produces the metric (2.7)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the Kruskal-coordinate transformation and conformal factor used to extract the $c_i(N)$ dependence."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the universal 2d CFT single-interval entanglement entropy formula used in (3.19)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides numerical checks supporting the use of the 2d entropy formula in higher-dimensional spacetimes, which the paper relies on to apply (3.19) to the full metric."},{"cited_title":"Karananas, Alex Kehagias, and John Taskas","cited_arxiv_id":null,"evidence_quote":"The linear-dilaton island result that the paper's two-interval expression reduces to when $\\Omega=1$ and $c=1$, serving as the comparison point for the LST calculation."}],"review_version":1}