{"id":"dbe88998-f25f-4892-bfa1-e94cd4e52d83","arxiv_id":"2507.16256","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Using a massive gravity formulation, the paper derives all-order leading-logarithmic T-bar-T corrections to two- and three-point CFT correlators, reproducing the known two-point result and obtaining a new closed form for the three-point function.","lead":"The paper computes all-order leading-logarithmic corrections to two- and three-point correlation functions in two-dimensional T-bar-T-deformed conformal field theories using a massive gravity description. The two-point result matches known formulas and the three-point result is presented as new.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unproven measure assumption and lack of an independent low-order check of the three-point formula leave the central new result conditional.","rationale":"The reader correctly identifies the omitted Jacobian and Weyl-mode factors in prescription (2.12) as the load-bearing assumption. My independent reading confirms the paper states but does not prove that these factors are leading-log subleading. The two-point reproducibility (Sec. 3) is not by itself dispositive: the two-point configuration has only one relative distance, and the Weyl/Jacobian factors could cancel there by symmetry while contributing for three points, where the cyclic structure and the ratio log(eps |x_BC|/(|x_AB| |x_AC|)) appear. There is a second, unaddressed technical point: the derivation of (4.7) depends on the momentum-space identity (3.8) applied separately to each pair via (4.3), which is only valid for generic non-integer conformal dimensions; the paper does not show that the special-case treatment in (3.12) passes through the three-point cyclic product. I do not see a more basic logical error: the Gaussian integration is standard, Eq. (4.4) is plausible, and the two-point result explicitly reproduces (2.13). I also credit the paper for signposting remaining issues and not overclaiming, since it admits that (4.7) generates subleading logs and only interprets the leading-log component as reliable. All this supports a CONDITIONAL verdict: the central new formula is likely right but too weakly checked to move to ACCEPT. The proposed concrete tests are the natural way to settle the matter. No ad hominem and no charge of circularity is intended; the concern is a specific, testable gap in justification.","tokens_in":13698,"tokens_out":1768,"duration_ms":16461,"concrete_test":"Expand both sides of (4.7) to second order in mu (i.e., O(mu^2) at leading log) for a configuration with generic unequal dimensions, using the known perturbative T-bar-T three-point leading-log corrections from direct conformal perturbation theory or from the random-geometry approach of [8] and [12]; if the O(mu^2) leading-log coefficient of (4.7) does not match the independent expansion, the omitted Jacobian/Weyl factors do matter at leading order. A simpler first check: switch on the Jacobian factors J(x_A) in (2.12) and compute the O(mu) three-point correction; if the leading log changes, the measure assumption fails.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's central new claim is the three-point leading-log formula, Eq. (4.7). Its derivation relies on the stripped-down prescription (2.12), which omits the Weyl-mode integration, the associated Jacobian factors J(x_A), and all non-Gaussian terms in SmG. The authors assert (Sec. 2.2, and again in Sec. 3.1.2) that these omissions do not change leading logarithms, but the assertion is not demonstrated by a direct calculation in the manuscript. The three-point derivation also uses the momentum-space representation (4.3), where the identity (3.8) requires generic (non-integer) dimensions; the integer-dimension continuation in Eq. (3.12) is described only for the two-point case and is not shown to be compatible with the three-point normal-ordering and cyclic structure in (4.7). The strongest concrete risk is that the omitted Jacobian and Weyl factors already contribute at leading-log order for the three-point correlator while canceling for the two-point function, so the empirical success of the two-point match would not protect the new three-point result. The paper itself labels the expression 'leading logarithmic' but then notes the differential operator in (4.7) also generates subleading logs, leaving some interpretive ambiguity about what is being claimed.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a prescription, Eq. (2.12), for computing $T\\bar{T}$-deformed correlators in a massive-gravity formulation: a Gaussian path integral over coordinate shifts $\\alpha_i(x_A)$ acting on the insertions of the undeformed CFT correlator, with the Weyl mode $\\phi$, the associated Jacobian factors, and non-Gaussian terms of the massive-gravity action omitted. Using this prescription the authors reproduce the known all-order leading-logarithmic two-point function, Eq. (2.13) (Cardy's result), at first, second, and all orders in the $T\\bar{T}$ coupling. They then extend the method to three-point functions and arrive at a new closed-form all-order expression, Eq. (4.7), written as a product of cyclic two-point correlators acted on by a normal-ordered exponential of a second-order differential operator. A nonperturbative completion of the two-point function is previewed in Eq. (3.13) and deferred to a companion paper. The main new claim is thus the three-point formula (4.7).","tokens_in":14025,"tokens_out":10219,"duration_ms":104396,"significance":"If correct, Eq. (4.7) provides a compact all-order expression for the leading-logarithmic three-point function in $T\\bar{T}$-deformed CFTs at finite coupling, which would be a new result and would demonstrate the utility of the massive-gravity framework as a finite-coupling generalization of the random-geometry approach. The two-point part is a useful and explicitly verified demonstration of the method: the authors reproduce a known external result with explicit analytic calculations, which is a genuine strength. However, the central three-point result rests on several unproven or compressed steps, and no independent low-order check of the three-point formula is provided. The paper is therefore best viewed as a promising but not yet fully established derivation of a new result; the significance is high conditional on filling these gaps.","major_comments":[{"comment":"","section":"Sec. 2.2 and Sec. 3.1.2, Eqs. (2.10)-(2.12)"},{"comment":"","section":"Sec. 4, Eqs. (4.3)-(4.7)"},{"comment":"","section":"Sec. 3.2, Eq. (3.12), and Sec. 4"},{"comment":"","section":"Sec. 4, text after Eq. (4.7)"}],"minor_comments":[{"comment":"","section":"Sec. 2.2, Eq. (2.12)"},{"comment":"","section":"Sec. 4, Eq. (4.7)"},{"comment":"","section":"Sec. 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is part of a closely knit series with the companion paper [11], and several of the strongest claims (nonperturbative completion, short-distance behavior) are deferred to that unpublished work. The refereed manuscript should stand on its own for the three-point formula; at present it does not, because the derivation is compressed and the measure assumption is unproven. I would encourage the editor to ask for a full derivation of (4.6) and for at least one independent low-order check of (4.7) before publication. There is also a potential overlap/competition with Ref. [12]; the 'Note added' is not a substitute for a quantitative comparison of the three-point results."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a solid step for the TbarT community. It reproduces the known all-order leading-log two-point function using the massive gravity framework, and it presents a genuinely new closed-form expression for the three-point function, Eq. (4.7). The Fourier-space derivation is clean, the geometric interpretation is attractive, and the authors are honest about what is leading-log and what is not. They also carefully distinguish their approach from Aharony-Barel's. The 2pt result matching Cardy at first, second, and all orders is real evidence that the framework is not empty. The soft spot is the measure. The prescription (2.12) drops the Weyl mode, the Jacobians, and the non-Gaussian terms in the massive gravity action, with the claim that these only affect subleading logarithms. That claim is plausible for the 2pt function, where the match works, but it is not demonstrated. For the 3pt function, there is no independent check - no low-order brute-force expansion, no argument that the omitted pieces cancel in the same way. The integer-dimension continuation is only spelled out for 2pt; the 3pt formula uses the same Fourier identity but the cyclic structure and normal-ordering might be more delicate. These are not fatal objections, but they make the new result conditional. A direct calculation of the leading-log contribution from the Jacobian, or a first-order brute-force check of (4.7), would settle the issue. The citation pattern is fine, and the authors acknowledge the relevant literature, including the competing [12]. The discussion section is speculative but clearly labeled as such. Who is this for? People working on TbarT deformations, random geometry, and holographic toy models. It is not a breakthrough for the whole field, but it is a useful technical advance that deserves referee time. I would accept this for peer review. The right request is a revision that either proves the measure simplifications or at least checks the 3pt formula at low order. Conditional acceptance, not outright rejection.","headline":"A useful finite-coupling extension of the random geometry approach that reproduces the known 2pt result and offers a new 3pt all-order leading-log formula, but the new result rests on an unproven measure simplification that needs explicit support.","tokens_in":711,"tokens_out":748,"would_cite":false,"duration_ms":27055,"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":"A Gaussian average over local coordinate shifts computes $T\\bar T$-deformed correlators to all orders, yielding a new three-point closed form.","keywords":["T\\bar T deformation","two-dimensional gravity","massive gravity","random geometry","correlation functions","leading logarithms","conformal field theory","all orders"],"falsifier":"Evaluate the full measure of (2.12), including the $\\phi$-integration and the Jacobian factors $J(x_A)=\\big[e^{2\\Phi(x_A)}\\det(\\delta_i^j+\\partial_j\\alpha^i(x_A))\\big]^{\\Delta_A/2}$, and compute the two-point correlator to second order in $\\mu$. If the $\\phi$ or Jacobian terms generate an $O(\\mu^2)$ contribution proportional to $\\ln^2(|x_{12}|/\\varepsilon)/|x_{12}|^{2\\Delta+4}$ with a coefficient different from $8\\Delta^2(\\Delta+1)^2/\\pi^2$, or any new leading-log structure, the all-order formula (2.13) is wrong; if they contribute only at subleading order, the leading-log claim stands. Applying the same test at $O(\\mu^2)$ to the three-point formula (4.7) would settle the new result.","tokens_in":13489,"feed_emoji":"📐","tokens_out":10101,"duration_ms":96561,"temperature":0.7,"pith_summary":"This paper is trying to establish that correlators of a two-dimensional conformal field theory deformed by the $T\\bar T$ operator can be computed at finite coupling by a simple prescription: take the undeformed CFT correlator, shift each operator insertion point by a dynamical coordinate transformation $\\alpha(x_A)$, and average over those shifts with a Gaussian weight coming from a two-dimensional massive-gravity action. The authors show that this prescription reproduces the known all-order leading-logarithmic series for the two-point function and produces a new compact all-order expression for the three-point function, Eq. (4.7), in which the deformed three-point correlator is a product of deformed two-point correlators of effective dimensions, acted on by an exponential differential operator. If correct, this turns the random-geometry picture of the $T\\bar T$ deformation into a practical finite-coupling computational tool and gives the first closed-form statement of the three-point leading-logarithmic structure.","feed_headline":"A 2D gravity picture computes all-order T-bar-T correlators","feed_subtitle":"The same random-coordinate average reproduces known two-point logs and yields a new three-point closed form.","key_machinery":"The carrying object is the massive-gravity action (2.1) with the metric parametrized as $ds^2 = e^{2\\Phi}\\delta_{ij}\\, d(x^i+\\alpha^i)\\,d(x^j+\\alpha^j)$, which splits the geometry into a Weyl mode $\\phi$ and diffeomorphism shifts $\\alpha^i$; with a preferred zweibein choice the action reduces to $\\frac{1}{\\mu}\\int(\\phi^2 + \\frac14 \\alpha^i \\Box \\alpha_i + O(\\alpha^3,\\alpha^2\\phi,\\ldots))$, the same Gaussian action that appeared in the random-geometry approach. The two-point propagator $\\Box^{-1}(x,x')=\\frac{1}{2\\pi}\\ln(|x-x'|/\\varepsilon)$ converts every Gaussian average into logarithms of insertion-point separations. The all-order evaluation relies on the momentum-space identity (3.8), which rewrites each CFT power $1/|x|^{2\\Delta}$ as a momentum integral; completing the square in $\\alpha$ exponentiates the momentum integral into $e^{(\\mu/\\pi)|k|^2\\ln(|x|/\\varepsilon)}$, so two- and three-point leading-log correlators reduce to products of two-point functions of lowered dimensions, producing the closed form (4.7).","core_discovery":"The central claim, on the paper's own terms, is that the prescription (2.12) — a path integral over diffeomorphism shifts $\\alpha^i(x)$ with Gaussian action $\\frac{1}{4\\mu}\\int \\alpha^i \\Box \\alpha_i$ and a specific preferred choice of zweibein parametrization — correctly captures the leading logarithmic corrections to $T\\bar T$-deformed correlators at every order in the coupling. For two points it yields the known series (2.13), verified explicitly at first and second order and then summed to all orders via a momentum-space identity. For three points it gives the new formula (4.7): the deformed correlator equals the CFT structure constant times, for each cyclic pair $(A,B,C)$, a normal-ordered exponential $\\exp\\!\\big(\\frac{\\mu}{2\\pi}\\ln\\frac{\\varepsilon |x_{BC}|}{|x_{AB}||x_{AC}|}\\frac{\\partial^2}{\\partial \\vec{x}_{AB}\\cdot \\partial \\vec{x}_{AC}}\\big)$ acting on a product of $T\\bar T$-deformed two-point functions of effective dimensions $\\delta_A = (\\Delta_B+\\Delta_C-\\Delta_A)/2$. Because the exponential operator also generates subleading logarithms, the paper conservatively presents only the leading-logarithmic part of (4.7) as reliable.","pith_inferences":["If (4.7) passes an explicit $O(\\mu^2)$ check, the same exponential-differential-operator structure should extend to $N$-point functions, with each pairwise channel contributing a deformed two-point factor; this is a direct testable next step.","The prescription suggests a stronger probabilistic reading of the deformation: the $T\\bar T$ coupling effectively averages the CFT over insertion-point displacements whose two-point variance grows as $\\mu\\ln(|x|/\\varepsilon)$; extracting the full distribution of $\\alpha(x_A)$ would make this interpretation precise.","The asserted innocuousness of the Jacobian, if confirmed elsewhere, would justify analogous omissions in stress-tensor correlator computations, where the Weyl mode is expected to contribute through the conformal anomaly.","The short-distance statistical incoherence may provide a quantitative bridge to braneworld or cutoff holography, where a fluctuating boundary geometry replaces a fixed radial cutoff; comparing (3.13) with holographic computations of such setups could sharpen the dictionary."],"forward_implications":["Two-point leading-logarithmic corrections are determined to all orders by (2.13), and the momentum-space derivation makes brute-force order-by-order expansion unnecessary.","Three-point leading-logarithmic corrections are captured in closed form by (4.7), the paper's claimed new result.","The framework extends the random-geometry computation of $T\\bar T$ correlators from infinitesimal to finite coupling; non-Gaussian terms in the massive-gravity action shift only subleading logarithms.","At short distances the correlators are typically suppressed, qualitatively similar to the short-distance behavior recently reported elsewhere, though with different quantitative details.","The nonperturbative completion previewed in (3.13) has the same large-distance asymptotic expansion as (2.13) and includes an instanton-like trans-series branch for positive $Z$."],"supporting_citations":[{"why":"Supplies the massive-gravity action (2.1) that defines the finite-coupling gravitational description used throughout.","marker":"[6]"},{"why":"Provides the random-geometry Gaussian action and structure that the prescription (2.12) is built on and must match.","marker":"[8]"},{"why":"Gives the known all-order leading-log two-point result (2.13), which the paper's two-point computation reproduces as its consistency check.","marker":"[9]"},{"why":"Offers the related dynamical-coordinate-transformation method whose two-point result is another point of comparison for the framework.","marker":"[10]"},{"why":"Underlies the diffusion-equation and random-geometry interpretation in (2.2)-(2.3) that justifies averaging over fluctuating geometries.","marker":"[17]"},{"why":"Provides the contrasting short-distance correlator results from a different gravitational description, against which the paper's findings are compared in the note added and Section 3.2.","marker":"[12]"}],"fun_headline_variants":["2D gravity yields all-order T-bar-T correlator logs","T-bar-T correlators from 2D gravity: all-order logs","New three-point logs in T-bar-T via gravity picture","Massive gravity computes T-bar-T log corrections","Gravity interpretation reproduces and extends T-bar-T logs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that omitting the Weyl-mode $\\phi$ integration and the Jacobian factors in the path-integral measure of (2.12) does not change the leading logarithmic contributions; the paper asserts this and verifies the two-point consequences, but does not display the explicit calculation of the omitted factors.","fun_headline_variants_meta":{"raw":{"variants":["2D gravity yields all-order T-bar-T correlator logs","T-bar-T correlators from 2D gravity: all-order logs","New three-point logs in T-bar-T via gravity picture","Massive gravity computes T-bar-T log corrections","Gravity interpretation reproduces and extends T-bar-T logs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000221,"raw_usage":{"total_tokens":1452,"prompt_tokens":952,"completion_tokens":500,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":568,"completion_tokens_details":{"reasoning_tokens":417}},"tokens_in":568,"tokens_out":500,"duration_ms":5387,"temperature":1.0,"reasoning_tokens":417,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T15:14:26.769628+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the full measure of (2.12), including the $\\phi$-integration and the Jacobian factors $J(x_A)=\\big[e^{2\\Phi(x_A)}\\det(\\delta_i^j+\\partial_j\\alpha^i(x_A))\\big]^{\\Delta_A/2}$, and compute the two-point correlator to second order in $\\mu$. If the $\\phi$ or Jacobian terms generate an $O(\\mu^2)$ contribution proportional to $\\ln^2(|x_{12}|/\\varepsilon)/|x_{12}|^{2\\Delta+4}$ with a coefficient different from $8\\Delta^2(\\Delta+1)^2/\\pi^2$, or any new leading-log structure, the all-order formula (2.13) is wrong; if they contribute only at subleading order, the leading-log claim stands. Applying the same test at $O(\\mu^2)$ to the three-point formula (4.7) would settle the new result.","supporting_citations":[],"review_version":1}