{"id":"40b2f595-a28f-478d-8a1e-1537c07ccac7","arxiv_id":"2608.07660","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"At large Im tau the superconformal index tends to 1, and the authors show that a Picard-Lefschetz analysis of a Lorentzian-inspired bulk integral enforces this by making all exponentially large black hole saddles irrelevant via Stokes phenomena.","lead":"This paper analyzes which gravitational saddle points contribute to the superconformal index in AdS/CFT. It argues that in a Lorentzian-inspired path integral, Stokes phenomena eliminate the black hole saddles that earlier work suggested would dominate at large chemical potentials.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The bulk conclusion that only thermal AdS contributes at large Im tau is conditional on the unproven ansatz (1.5) for rotating AdS5 black holes and on the hand-imposed truncation to truly-BPS saddles in section 3.2; neither is derived, so the Stokes-based exclusion of the (1.4) saddle is not yet…","rationale":"I read the paper as a careful and transparent attempt to resolve a real tension in holographic index computations. The exact CFT-side argument in Appendix A that the index approaches 1 at fixed N and large Im tau is clean and parameter-free; it independently rules out contributions exponentially large in N. The novel and load-bearing part is the bulk mechanism: the Lorentzian-motivated ansatz (1.5), the reduction to the BPS-only ansatz (3.1), and the by-hand truncation to truly-BPS saddles in section 3.2. The paper explicitly flags that the derivation of (1.5) for rotating higher-dimensional black holes is not available, and section 7 presents the truncation as a conjecture. The reader's weakest_assumption identifies exactly these points, and I agree that they make the bulk conclusion conditional rather than established. My stress-test does not find an internal inconsistency or a stronger objection; the concern is that the central bulk claim is only as secure as these acknowledged assumptions. The proposed one-loop fermion determinant check would directly test the truncation step, which is the most concrete place where the argument could fail. Since the reader already assigned CONDITIONAL on essentially these grounds, my recommendation is UNCHANGED.","tokens_in":51244,"tokens_out":5180,"duration_ms":52092,"concrete_test":"Compute the one-loop fermion determinant around the finite-beta non-BPS saddle of section 4.2 that tends to t=(4+3i)/5 (section 5.1), at a representative tau inside the red subregion of Figure 6 with 2 tau_n - 3 Delta_{n'} = 1, and check whether the determinant vanishes identically as required to kill the only-asymptotically-BPS contribution. If it does not vanish, the section 3.2 truncation is unjustified and the Stokes-based exclusion of (1.4) at large Im tau is not established; if it does vanish, this particular concern is settled.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central bulk claim—that the exponentially large saddle (1.4) is removed by Stokes phenomena and only thermal AdS survives at large Im tau—rests on two unproven steps. First, section 1 states that the Lorentzian-motivated ansatz (1.5) has been derived only for Maxwell charges and for angular momentum in 2+1 dimensions; for AdS5 rotation it is taken 'as motivation' without a derivation. If the correct contour or measure differs from (1.5), the Picard-Lefschetz intersection numbers computed in section 5 can change, so the disappearance of t+ at large Im tau is not guaranteed for the actual gravitational path integral. Second, section 3.2 proposes to truncate the sum over saddles to 'truly-BPS' saddles by hand, with only a weaker consistency condition checked in the restricted model; section 7 calls the resulting prescription a conjecture supported by Occam's razor. The only-asymptotically-BPS saddle at t=(4+3i)/5 is found to contribute to the BPS-only ansatz in a subregion of Figure 6, and its removal depends on this unproven truncation. The analysis is also limited to equal charges and SO(6)-homogeneous internal spaces, excluding D3/M5-brane effects acknowledged in section 7. The CFT bound in Appendix A independently gives I to 1 at fixed N and large Im tau, so the abstract's statement about the index is robust; but the paper's bulk explanation—the advertised Stokes mechanism—is not yet derived, only modeled.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the large-Im tau behavior of the superconformal index of N=4 SU(N) SYM and its ABJM analogue, arguing that the bulk gravitational path integral, when defined through a Lorentzian-motivated ansatz, receives only thermal AdS contributions at sufficiently large Im tau. The authors reduce the bulk problem to a one-dimensional BPS integral, perform a Picard-Lefschetz analysis of its saddles, and show that the exponentially large black hole saddle (1.4) is removed by a Stokes phenomenon. They also present a CFT-side derivation showing I approaches 1 at fixed N for large Im tau.","tokens_in":51518,"tokens_out":3320,"duration_ms":32605,"significance":"The paper contains several strong technical components: a clean analytic derivation of the CFT index asymptotics in Appendix A (Eq. (A.8)), an explicit reduction to a one-dimensional BPS integral in Section 5, and careful numerical Picard-Lefschetz flow computations in Figures 3-5 and Appendix C.2 that substantiate the Stokes analysis. If the bulk ansatz and truncation were derived from first principles, the conclusion that previously proposed black hole saddles are absent at large Im tau would be an important resolution of a known tension. The authors are transparent about the conditional status of these ingredients, which is a credit to the manuscript.","major_comments":[{"comment":"The central bulk claim is conditional on the Lorentzian-motivated ansatz (1.5), and the text explicitly states that the required derivation exists only for Maxwell charges and for angular momentum in 2+1 dimensions, while for AdS5 rotation the formula is taken 'as motivation.' The Picard-Lefschetz analysis in Section 5 computes saddle relevance for this specific reduced integral, so if the correct gravitational contour or measure differs from (1.5), the Stokes removal of the t+ saddle at large Im tau is not established for the actual index. The CFT bound in Appendix A is independent and robust, but the advertised bulk mechanism needs either a derivation of (1.5) for rotating AdS5 or a controlled test in a setting where the Lorentzian path integral can be computed independently.","section":"Section 1, Eq. (1.5)"},{"comment":"The truncation of the saddle sum to 'truly-BPS' saddles is a conjecture; the paper checks only a weaker consistency condition, and the only-asymptotically-BPS saddle at t=(4+3i)/5 is found to contribute in the red subregion of Figure 6 and is then removed by hand. Since this saddle is the large-beta limit of a non-BPS saddle with a fermion zero mode, its removal is plausible, but the weaker condition does not prove that the truncated sum is the semiclassical expansion of an integral representing the true index. Without a derivation from fermionic localization, the conclusion that only thermal AdS survives at large Im tau does not follow for the full index.","section":"Section 3.2 and Section 7"},{"comment":"The analysis is restricted to tau=sigma, equal charges, equal angular momenta, and SO(6)-homogeneous internal spaces, and D3/M5-brane effects are not included. The abstract's statement that black hole saddles 'cannot contribute' is broader than what is demonstrated; the demonstrated statement concerns the restricted BPS ansatz. This is an acknowledged limitation, but it is load-bearing for the claimed exclusion of previously suggested saddles, so the scope of the central claim should be narrowed or the missing effects must be addressed.","section":"Section 4 and Section 7"}],"minor_comments":[{"comment":"'In section 3 we will find values...' should refer to Section 5 (or Section 4.2), since the relevant only-asymptotically-BPS saddle is analyzed there.","section":"Section 3.2, first sentence"},{"comment":"The color labels 'lavender' and 'red' do not clearly correspond to the printed shading; please add a legend or hatching labels for accessibility.","section":"Figure 6 caption"},{"comment":"There are several typographical errors, including 'Stoke's transitions' in Section 5.1 and inconsistent use of 'ReSE = 0' in Section 5.3 where the BPS action is intended; please proofread.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"This is a serious and carefully written paper, but the headline conclusion rests on two unproven pillars: the rotating version of the Lorentzian ansatz and the truncation to truly-BPS saddles. The authors are unusually explicit about these gaps, and the technical analysis within the model is convincing. I recommend major revision rather than rejection, because the gaps are clearly identified and may be addressable, but the current framing overstates what has been established for the actual index."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line first. This is a serious paper that takes on a real, well-known tension in the holographic superconformal index, and the abstract's central index claim actually holds up: the CFT-side bound is robust and independent of the bulk machinery. The bulk mechanism that explains it, however, is modeled rather than derived, and the paper says so explicitly.\n\nWhat's new and good. Appendix A derives the finite-N asymptotics I = 1 + O(e^{-8 pi Im tau / 3}) analytically, from the single-letter index plus a one-line Weingarten argument. That is clean, parameter-free, and not in the earlier literature; reference [6] only had it numerically. On the bulk side, the reduction to a one-dimensional BPS integral, the explicit reduced action (5.13), and the Picard-Lefschetz analysis with the phase diagrams in Figures 2 and 6 are careful and concrete. The claim that the exponentially large saddle (1.4) is removed by Stokes phenomena at large Im tau is genuinely new, against the suggestions of [5,7]. The finite-beta numerics of Appendix C, where two-complex-dimensional flows reproduce the intersection numbers of the one-dimensional reduction, are a meaningful consistency check, and the bisection method is described well enough to reproduce.\n\nSoft spots, in proportion. Everything in the bulk conclusion hangs from the ansatz (1.5), which is derived for Maxwell charges and for angular momentum in 2+1 dimensions but is taken as motivation for AdS5 rotation; the paper says this in section 1. If the correct contour prescription differs, the section 5 intersection numbers could change. Second, the truncation to truly-BPS saddles in section 3.2 is hand-imposed, and section 7 calls it a conjecture. In fairness, the paper does not just wave its hands: it requires that removed saddles never catalyze Stokes transitions for retained ones, and checks that weaker condition in both the AdS5 and AdS4 systems. But the only-asymptotically-BPS saddle at t = (4+3i)/5 does contribute to the BPS-only ansatz in a subregion of Figure 6, and its removal depends on the unproven truncation. Equal charges, homogeneous internal spaces, and the neglect of D3/M5-brane effects further narrow the scope, as the authors acknowledge. The numerical Stokes analyses also come without shipped code or data, so some flow statements are only as checkable as the text's figures and description.\n\nThe citation pattern is on point: the Bethe-ansatz, matrix-model, and Lorentzian path-integral literature is engaged where it matters, and the self-citations point to the prior work the analysis actually builds on. This is worth reading for anyone working on holographic indices or the Lorentzian path integral program, and it deserves a serious referee. The gaps are explicit, the CFT result is solid, and the bulk analysis is a well-executed proposal whose load-bearing assumptions are named.","headline":"A careful, honest paper that resolves the tension on the CFT side with a clean analytic bound, but whose advertised bulk mechanism rests on an unproven ansatz and a hand-imposed truncation, both acknowledged.","tokens_in":52136,"tokens_out":7039,"would_cite":true,"duration_ms":59455,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"At sufficiently large ${\\rm Im}\\,\\tau$ the superconformal index is dominated by the thermal AdS saddle, so all black hole saddles with exponentially large $N$ contributions cannot contribute.","keywords":["superconformal index","AdS/CFT correspondence","black hole saddles","Picard-Lefschetz theory","Stokes phenomena","N=4 supersymmetric Yang-Mills","ABJM theory","Lorentzian path integral"],"falsifier":"Compute the ${\\mathcal N}=4$ SU($N$) matrix integral (or its Bethe-ansatz form) at large $N$ for a value of $\\tau$ outside the predicted Stokes region, e.g. $\\tau=0.3+0.4i$, and look for a term proportional to $\\exp(-N^2 i\\pi(2\\tau-1)^3/(27\\tau^2))$ with a non-vanishing coefficient; if such a term survives, the claimed Stokes removal is wrong. A complementary check is to perform the gravitational path integral with an independent contour prescription and test whether the black hole saddle's Lefschetz thimble intersects the integration contour at that $\\tau$.","tokens_in":50941,"feed_emoji":"🕳️","tokens_out":18086,"duration_ms":154406,"temperature":0.7,"pith_summary":"This paper resolves a tension inside AdS/CFT. The superconformal index of ${\\mathcal N}=4$ supersymmetric Yang-Mills theory is known from its matrix-integral definition to tend rapidly to 1 when the fugacity $\\tau$ has a large imaginary part, for any rank $N$; yet earlier bulk analyses had proposed a black hole saddle whose contribution grows exponentially with $N$ in exactly that regime. The authors argue that the bulk gravitational path integral should be defined through a Lorentzian-motivated ansatz that sums over real stationary black holes with codimension-2 conical defects, and that fermion zero modes further restrict the relevant configurations to the BPS locus. Using Picard-Lefschetz thimble analysis on the resulting one-dimensional integral, they show that the previously proposed exponential saddle contributes only inside a bounded region of the $\\tau$-plane and is turned off by a Stokes phenomenon at sufficiently large ${\\rm Im}\\,\\tau$. The same analysis for the AdS$_4$/ABJM index produces the same conclusion: in the large-${\\rm Im}\\,\\tau$ regime the only remaining saddle is thermal AdS, so the index is of order one, consistent with the CFT side. If correct, this establishes which bulk saddles actually appear in these index computations and explains why the earlier exponentially large contributions are absent.","feed_headline":"Superconformal index drops to 1 at large Im tau","feed_subtitle":"Previously proposed black hole saddles are removed by Stokes phenomena, leaving thermal AdS to dominate.","key_machinery":"The load-bearing object is the Lorentzian-motivated ansatz (1.5), an integral over the area $A$, angular momenta $J_1,J_2$, and R-charges $Q_1,Q_2,Q_3$ of real stationary Lorentz-signature black holes with codimension-2 singularities, weighted by $e^{A/4G_5} e^{-\\beta(E-\\Omega_1J_1-\\Omega_2J_2 - \\frac12\\sum_i \\Phi_i Q_i)}$ times shift phases. Fermion zero modes are argued to reduce this, in the $\\beta\\to\\infty$ limit, to the BPS-only ansatz (3.1), which inserts a delta-function $\\delta(E-2J-\\frac32 Q)$ and integrates over the one-dimensional locus of real extremal BPS black holes. In the equal-charge, equal-angular-momentum case for AdS$_5$ the BPS locus is parametrized by $a\\in[0,1)$, mapped to $t\\in[1,2+\\sqrt3)$; the reduced integral is one-dimensional with an action $\\tilde S_{\\rm BPS}(t;\\tau)$ whose saddles include $t_\\pm$, the thermal-AdS endpoint $t=1$, and $\\tau$-independent points $t=\\pm i$ and $t=(4+3i)/5$. Its contribution structure is decided by Picard-Lefschetz thimbles—the upward-flow cycles from each saddle and their intersection with the real contour—and by Stokes phenomena, where an ascent thimble jumps discontinuously when it encounters another saddle that contributes to the integral. The catalog of which saddles catalyze which transitions is what turns the black hole contribution on and off as $\\tau$ varies.","core_discovery":"The central claim, stated on the paper's own terms, is that at large ${\\rm Im}\\,\\tau = {\\rm Im}\\,\\sigma$ the superconformal index rapidly approaches 1 at all values of $N$, so no black hole saddle with a contribution exponentially large in $N$ can contribute to it. For the equal-charge truncation of the AdS$_5$ computation, the paper shows that the saddle with action $\\exp(-N^2 i\\pi(2\\tau-1)^3/(27\\tau^2))$—the 'unshifted' black hole saddle—is relevant only for a limited range of $\\tau$, and that above a finite ${\\rm Im}\\,\\tau$ a Stokes transition catalyzed by the thermal AdS endpoint removes it from the semiclassical expansion. The same mechanism applies to the shifted sectors and to orbifold/quotient saddles, and the AdS$_4$ equal-charge analysis of the ABJM index yields an analogous phase diagram in which a truly-BPS black hole saddle contributes only inside $|\\tau|<1$. In both cases the infinite sums over shifted sectors collapse: at most one black hole saddle contributes for AdS$_5$, at most two for AdS$_4$, and at sufficiently large ${\\rm Im}\\,\\tau$ only the thermal AdS endpoint remains. The authors further propose that saddles which are only asymptotically BPS must be dropped from the sum, and verify the consistency condition that those dropped saddles never catalyze Stokes transitions for the truly BPS ones.","pith_inferences":["If the Lorentzian ansatz is the right definition of the gravitational path integral, the same Stokes phenomenon should select saddles in other supersymmetric index computations, including refined indices with unequal fugacities, because the mechanism is the structure of thimbles rather than the equal-charge simplification.","A sharp, testable prediction the paper does not spell out is that the turn-on of the exponentially large saddles is a non-analyticity in $\\tau$ located on the predicted Stokes curves of the phase diagram; evaluating the large-$N$ matrix integral near those curves on the CFT side could directly confirm or refute the bulk picture.","For correlation functions whose inserted operators soak up fermion zero modes, the finite-$\\beta$ analysis in appendix B shows that only-asymptotically-BPS saddles can catalyze Stokes transitions for truly-BPS saddles, so the main-text truncation recipe should not be applied to such supersymmetry-breaking insertions.","A general selection rule suggested by the paper's logic is that a complex saddle contributes to a supersymmetric index only if it is the large-$\\beta$ limit of a finite-$\\beta$ saddle that remains BPS; applying this rule would remove many candidate saddles in other gravitational index computations before any contour analysis."],"forward_implications":["In the large-${\\rm Im}\\,\\tau$ regime the index is of order $N^0$; the thermal AdS endpoint is the only contributing bulk saddle, so the previously proposed exponentially large black hole contribution is absent there.","At any fixed $\\tau$ at most one AdS$_5$ shifted-sector black hole saddle, and at most two in AdS$_4$, can contribute; the infinite sums over shifts and over orbifold sectors reduce to order-one thermal-AdS coefficients plus finitely many saddles.","Supersymmetric orbifold saddles inherit the Stokes structure of their covering black hole with the $\\tau$-region scaled by $1/m$, so at sufficiently large ${\\rm Im}\\,\\tau$ they too disappear from the index.","The AdS$_5$ matrix-integral asymptotics, ${\\mathcal I}(\\tau)=1+O(e^{-8\\pi\\,{\\rm Im}\\,\\tau/3})$ at fixed $N$, are the CFT-side fingerprint of the bulk conclusion; the paper predicts no exponentially-in-$N$ correction to this behavior.","The consistency condition that only-asymptotically-BPS saddles never catalyze Stokes transitions for truly-BPS saddles is satisfied by both systems, supporting the proposal to truncate the saddle sum to truly-BPS saddles."],"supporting_citations":[{"why":"Defines the ${\\mathcal N}=4$ SYM superconformal index and establishes its independence of $\\beta$ at fixed $\\tau,\\sigma,\\vec\\Delta$.","marker":"[1, 2]"},{"why":"Gives a Bethe-ansatz computation of the index from which the exponentially large black hole saddle contribution arises, the main target of the paper's exclusion argument.","marker":"[5]"},{"why":"Provides the CFT-side numerical observation that the index approaches unity in the limit that corresponds to large ${\\rm Im}\\,\\tau$, the behavior the bulk analysis reproduces.","marker":"[6]"},{"why":"Identifies the additional bulk saddles, including orbifolds, whose semiclassical contributions the paper re-analyzes with Picard-Lefschetz methods to determine when they actually contribute.","marker":"[7]"},{"why":"Supplies the Lorentzian path-integral prescription with codimension-2 singularities on which the ansatz (1.5) is based.","marker":"[34]"},{"why":"Derives the ansatz for Maxwell charges and for angular momentum in 2+1 dimensions, the motivation for writing the corresponding formulae in higher dimensions.","marker":"[41]"},{"why":"Establishes that fermion zero modes remove non-BPS saddles from supersymmetric indices, the input that selects the BPS-only ansatz (3.1).","marker":"[47]"},{"why":"Provides the AdS$_5$ black hole solutions and their thermodynamic potentials used to build the BPS locus and the reduced one-dimensional action.","marker":"[58]"},{"why":"Gives the AdS$_4$/ABJM analog of the bulk saddle analysis whose complex saddles and shifted sectors the paper re-examines.","marker":"[61]"}],"fun_headline_variants":["Index approaches 1 at large imaginary tau, killing black hole saddles","Stokes phenomena drop black hole saddles from superconformal index","Superconformal index trivializes at large Im tau, excluding black holes","No black hole saddles when Im tau is large: index approaches 1","Large Im tau forces superconformal index to 1, dropping black hole saddles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion depends on the assumption that the bulk gravitational path integral is correctly captured by the Lorentzian-motivated ansatz—an integral over real stationary black holes with codimension-2 singularities—together with the hand-imposed restriction of the sum over saddles to those that are truly BPS; if the correct contour prescription or the truncation differs, the large-${\\rm Im}\\,\\tau$ dominance of thermal AdS need not follow for the actual index.","fun_headline_variants_meta":{"raw":{"variants":["Index approaches 1 at large imaginary tau, killing black hole saddles","Stokes phenomena drop black hole saddles from superconformal index","Superconformal index trivializes at large Im tau, excluding black holes","No black hole saddles when Im tau is large: index approaches 1","Large Im tau forces superconformal index to 1, dropping black hole saddles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001123,"raw_usage":{"total_tokens":4734,"prompt_tokens":1073,"completion_tokens":3661,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":689,"completion_tokens_details":{"reasoning_tokens":3562}},"tokens_in":689,"tokens_out":3661,"duration_ms":22880,"temperature":1.0,"reasoning_tokens":3562,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T00:26:32.735472+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the ${\\mathcal N}=4$ SU($N$) matrix integral (or its Bethe-ansatz form) at large $N$ for a value of $\\tau$ outside the predicted Stokes region, e.g. $\\tau=0.3+0.4i$, and look for a term proportional to $\\exp(-N^2 i\\pi(2\\tau-1)^3/(27\\tau^2))$ with a non-vanishing coefficient; if such a term survives, the claimed Stokes removal is wrong. A complementary check is to perform the gravitational path integral with an independent contour prescription and test whether the black hole saddle's Lefschetz thimble intersects the integration contour at that $\\tau$.","supporting_citations":[{"cited_title":"A gravity interpretation for the complex Euclidean saddles of the ABJM index","cited_arxiv_id":"2605.00987","evidence_quote":"Gives the AdS$_4$/ABJM analog of the bulk saddle analysis whose complex saddles and shifted sectors the paper re-examines."}],"review_version":1}