{"id":"57144b40-203d-4dd9-9916-da8ae2402887","arxiv_id":"2508.21660","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Using CONTUR reinterpretation of LHC data, the authors derive 2 sigma bounds on ALP-top and ALP-gluon couplings, with fa/c_t above about 200 GeV and fa/c_gluon above about 5 TeV.","lead":"This paper uses existing LHC measurements to constrain axion-like particles that couple mainly to top quarks and gluons. It shows that loop effects let otherwise hard-to-see ALPs show up in precision data, yielding bounds on the ALP coupling plane.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline limits are quoted without the sqrt(s_hat) < f_a EFT-validity cut that Sec.","rationale":"The paper is a serious, carefully written reinterpretation study. The CONTUR methodology, the loop-corrected effective couplings, and the explicit virtual-exchange-only control case in Sec. 4.3 are all real strengths. I considered whether the invisible-ALP assumption identified by the reader is the most load-bearing issue; the virtual-only option weakens but does not remove the bounds, so the central claim does not stand or fall on that assumption alone. The more serious issue is internal to the EFT logic: Sec. 5 establishes sqrt(s_hat) < f_a as the validity criterion, then presents the main exclusion contours without applying it. The quoted boundary f_a/c0_t ~ 200 GeV is exactly where LHC partonic energies exceed f_a for order-one couplings, so the limit as stated is not yet an EFT-valid statement. This does not make the paper's approach wrong, but it means the headline quantitative limits need a validity cut or at least a validity-truncated contour before they can be interpreted as constraints on f_a. Since the reader already assigned CONDITIONAL and also noticed the Sec. 5 awkwardness, my assessment leaves the verdict unchanged; it sharpens the specific condition that should be satisfied in a revision. I did not find evidence of internal mathematical inconsistency in the loop formulas, and the lack of shipped code, while limiting reproducibility, is not itself a correctness flaw.","tokens_in":22183,"tokens_out":12287,"duration_ms":118980,"concrete_test":"Re-run the CONTUR scan with a per-event EFT-validity cut, retaining only simulated signal events with sqrt(s_hat) < f_a, where f_a is inferred at each parameter point from the plotted ratio c/f_a (e.g. take c0_t = 1 along the c0_t axis and set f_a = c0_t / (c0_t/f_a)). Regenerate Figs. 5 and 12 and compare the 2-sigma excluded area with and without the cut. If the excluded area shrinks appreciably, or if the boundary point f_a/c0_t ~ 200 GeV with c0_t ~ 1 is driven by events with sqrt(s_hat) > f_a, then the headline limits must be re-presented as validity-truncated contours.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim maps LHC measurements onto the ALP EFT parameter plane and quotes 2-sigma limits such as f_a/c0_t >~ 200 GeV and f_a/c0_tildeG >~ 5 TeV. For this mapping to be meaningful, the EFT expansion in powers of E/f_a must be valid, i.e. sqrt(s_hat) < f_a, as the paper itself states in Sec. 5. However, Sec. 5 also explicitly says that the limits in Figs. 5 and 12 were obtained without imposing any such cut: 'no cut has been applied on the partonic center-of-mass energy sqrt(s_hat)'. The quoted boundary f_a/c0_t >~ 200 GeV is precisely the dangerous regime: for order-one c0_t, this corresponds to f_a around 200 GeV, while the sensitive bins shown in Fig. 7b and Fig. 11b extend above 500 GeV to more than 1 TeV. Events at these momenta violate the paper's own validity condition, so the signal predictions used to exclude those points are outside the domain of the EFT. The statement that 'the kinematic regions where the ALP signal manifests most prominently lie in the few hundred GeV range' is asserted but not demonstrated for the actual exclusion boundary, and it is in tension with the high-pT bins that visibly drive the exclusions. Thus the contours as plotted are not yet shown to be valid EFT constraints; a validity-truncated contour could be substantially smaller, especially along the c0_t axis. This is a self-flagged limitation in Sec. 5, but it is left unresolved in the presentation of the main limits.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses the CONTUR framework to reinterpret a broad set of 13 TeV LHC measurements (dijet, lepton+jets and dilepton top-pair, all-hadronic top-pair, photon+jets, Z+jets, and missing-energy+jets final states) in an ALP EFT in which only the top-quark and gluon couplings are non-zero at tree level. One-loop corrections from Ref. [46] are included, mapping the bare couplings (c0_t, c0_tildeG) to effective couplings ceff_t and ceff_agg, and inducing additional couplings to light fermions and electroweak gauge bosons. Exclusion contours are presented in the (ceff_agg/fa, ceff_t/fa) and (c0_tildeG/fa, c0_t/fa) planes for ma = 1 GeV and Lambda = 10 TeV, under the stated assumption that final-state ALPs are invisible at detector level; Sec. 4.3 provides a weaker alternative with only virtual ALP exchange. The paper also sketches an SO(6)/SO(5) composite-Higgs UV completion and discusses EFT validity in Sec. 5.","tokens_in":22466,"tokens_out":9366,"duration_ms":88902,"significance":"If the constraints survive scrutiny, the paper demonstrates that existing public LHC measurements can probe ALP couplings to tops and gluons, including regions where the ALP is effectively invisible. The analysis is transparent about its main assumption, provides a conservative virtual-only variant, includes momentum-dependent one-loop corrections, and relies on external data, so the central constraints are not circular. The main limitation is the unresolved EFT-validity issue described below, which affects the quoted headline limits.","major_comments":[{"comment":"The headline exclusions are not validity-truncated. Section 5 states that no cut has been applied on the partonic center-of-mass energy sqrt(s_hat), yet the quoted 2sigma limit fa/c0_t >~ 200 GeV places the exclusion boundary at fa ~ 200 GeV for order-one c0_t. The histograms that drive the exclusions in Fig. 7b (top pT up to 1600 GeV) and Fig. 11b (Z pT up to 1000 GeV) extend well above this scale, so the signal predictions used to exclude those points violate the paper's own criterion sqrt(s_hat) < fa. The assertion that the sensitive kinematic regions lie in the few-hundred-GeV range is asserted but not demonstrated for the actual exclusion boundary, and it is in tension with those figures. I ask that the contours be re-derived with a validity cut, or equivalently that parameter regions with fa below the characteristic scale of the sensitive bins be masked, and that the quoted limits be restated for the valid region.","section":"Sec. 5; Figs. 5 and 12"},{"comment":"The virtual-ALP-only limits, presented as the conservative alternative, are obtained under the same no-cut assumption as Fig. 5. Even with ALP production diagrams removed, the off-shell ttbar and dijet processes receive contributions from the same high-pT bins, so the validity concern carries over. The same truncation should be applied to Fig. 12, and the text should state whether the weaker bounds survive in the valid region.","section":"Sec. 4.3 and Fig. 12"},{"comment":"The conversion from the effective-coupling exclusions to the bare-coupling plane depends on the arbitrary cutoff Lambda, which is fixed to 10 TeV. Because the headline limits are quoted in the (c0_t, c0_tildeG) plane, the paper should quantify how those limits shift under a reasonable variation of Lambda (for example 5-20 TeV); the logarithmic dependence is expected to be mild, but it should be demonstrated rather than assumed.","section":"Sec. 2.2, Eq. (2.22)"}],"minor_comments":[{"comment":"The phrase 'leading to to an increased cross section' contains a duplicated 'to'; it should read 'leading to an increased cross section'.","section":"Sec. 4.2"},{"comment":"The word 'particuarly' should be spelled 'particularly'.","section":"Sec. 4.2"},{"comment":"The statement that the most constraining regions arise from events with p >~ 200 GeV is vague; please specify whether p refers to the transverse momentum of the leading jet, the top quark, or the partonic center-of-mass energy, since the validity condition is stated for sqrt(s_hat).","section":"Sec. 5"},{"comment":"The word 'model-independent' is used for bounds that depend on the invisible-ALP assumption and on the choice Lambda = 10 TeV; please add qualifying phrases such as 'under the stated assumptions'.","section":"Abstract and Sec. 6"},{"comment":"The caption refers the reader to the legend of Fig. 4, but the two figures have different axis ranges and are not contiguous; repeating the legend or naming the pools in the caption would improve readability.","section":"Fig. 5"},{"comment":"The master formula for the effective fermion coupling is presented without the explicit loop functions Di, which are only given in Ref. [46]; since this equation is central to the analysis, please consider listing the definitions or at least stating the sign conventions used.","section":"Eq. (2.21)"}],"recommendation":"major_revision","confidential_remarks":"The main issue in my view is the unresolved EFT-validity truncation: the quoted limits are derived without imposing sqrt(s_hat) < fa, and the low-fa region of the exclusion boundary is precisely where the paper's own validity criterion fails. This is fixable by re-deriving the contours with a validity cut and restating the headline numbers. The overlap with Ref. [46] is not a circularity problem because the experimental constraints come from external LHC data, but the authors should avoid overstating model-independence given the invisible-ALP and Lambda = 10 TeV assumptions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the first LHC-data constraint on the low-mass ALP-gluon coupling and its interplay with the ALP-top coupling. The authors include loop-induced couplings, use a wide range of public measurements through CONTUR, and are transparent about the invisible-ALP assumption by providing a virtual-exchange-only version in Sec. 4.3 that weakens but doesn't remove the bounds. That's responsible practice, and the dijet and top-pair exclusions look plausible.\n\nThe soft spot that matters most is the EFT validity issue. Sec. 5 states the condition sqrt(s_hat) < f_a, then says the contours in Figs. 5 and 12 were obtained without applying any such cut. The quoted boundary f_a/c0_t ~ 200 GeV is exactly the dangerous regime: for order-one c0_t that means f_a around 200 GeV, while the sensitive bins in Figs. 7b and 11b extend above 500 GeV, often past 1 TeV. The claim that the signal sits in the few-hundred-GeV range is asserted, not demonstrated for the actual exclusion boundary, and the high-pT bins visibly drive those exclusions. A validity-truncated contour would be smaller, especially along the c0_t axis. The authors flag this limitation, but they leave it unresolved in the headline limits. For a paper whose central result is the exclusion contour, that's a real gap, not a nitpick.\n\nA couple of smaller things. The invisible-ALP assumption is a genuine model dependence, but they handle it honestly, so I don't hold it against them. Fig. 7b uses a distribution where the SM baseline already has p = 0.09; the ALP makes it worse, but a poorly described baseline is a shaky foundation for a strong exclusion. And there's an internal inconsistency in Sec. 5: the text says order-one values of both c0_t and c0_tildeG remain compatible for f_a >~ 1 TeV, but the quoted limit f_a/c0_tildeG > 5 TeV implies c0_tildeG < 0.2 for f_a = 1 TeV. Probably a typo, but it should be fixed.\n\nWho is this for? People working on ALP phenomenology or on EFT reinterpretation of LHC data. It fills a gap and uses a sound, if established, methodology. It deserves a serious referee, not a desk reject. The referee should ask for contour plots with the sqrt(s_hat) < f_a cut applied, and for the text inconsistency to be cleaned up. I'd engage with it, but I wouldn't quote the current contours without the validity caveat.","headline":"A competent CONTUR reinterpretation that delivers the first real-data constraints on the low-mass ALP-gluon coupling and its interplay with the top coupling, but the headline contours are drawn without applying the paper's own EFT-validity cut, so the limits as plotted are not yet trustworthy.","tokens_in":23056,"tokens_out":5476,"would_cite":true,"duration_ms":48367,"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":"Run 2 LHC data, re-analyzed with loop-corrected effective field theory, exclude much of the axion–top–gluon coupling plane.","keywords":["axion-like particles","ALP-top coupling","ALP-gluon coupling","effective field theory","loop corrections","LHC reinterpretation","top quark","missing energy"],"falsifier":"An independent calculation of the one-loop mixing coefficient that produces $c^\\mathrm{eff}_t$ from $c_{\\tilde G}^0$ (the factor multiplying $c_{\\tilde G}^0$ in Eq. (2.22), shown to be about 0.4–0.7 in the relevant energy range) that disagrees with the paper's result would invalidate the mapping between bare and effective couplings and therefore the quoted exclusion limits.","tokens_in":21967,"feed_emoji":"🎯","tokens_out":13128,"duration_ms":106362,"temperature":0.7,"pith_summary":"ALPs that couple at tree level only to top quarks and gluons are hard to see directly, but this paper shows that one-loop radiative effects generate couplings to photons, Z bosons, and lighter fermions, so ordinary LHC measurements become probes of the ALP. Reinterpreting a wide range of Run 2 data—top-pair, dijet, photon+jet, and Z+jet distributions—the paper finds 2σ exclusions in the plane of the bare couplings: $f_a/c_t^0 \\gtrsim 200$ GeV and $f_a/c_{\\tilde G}^0 \\gtrsim 5$ TeV, with $|c_{\\tilde G}^0| \\gtrsim 0.5$ excluded in dijet final states and $|c_t^0| \\gtrsim 2$ excluded in semileptonic top-pair final states. The limits depend on the UV origin of the couplings, and the paper shows they already touch well-motivated UV scenarios such as composite Higgs models. The point is that loop-aware reinterpretation of existing measurements can probe otherwise elusive ALP scenarios, with sensitivity expected to improve at the HL-LHC.","feed_headline":"Run 2 data already exclude much of the axion–top–gluon plane","feed_subtitle":"Loop corrections turn ordinary top-pair and dijet measurements into constraints on ALP couplings.","key_machinery":"The key machinery is the set of one-loop corrected effective couplings that connect the two bare ALP couplings to the full low-energy ALP Lagrangian. In particular, the effective ALP–gluon coupling is $c^\\mathrm{eff}_{agg} = c_{\\tilde G}^0 - (\\alpha_s/8\\pi)\\, c_t^0\\, B_1(4m_t^2/p^2)$ and the effective ALP–top coupling is $c^\\mathrm{eff}_t = c_t^0 + \\frac{4\\alpha_s}{3\\pi} [3\\log(\\Lambda^2/m_t^2) - 4 - \\frac{2\\pi^2}{3} - \\frac{1}{2}(\\log(m_t^2/p^2)+i\\pi)^2]\\, c_{\\tilde G}^0$, with the loop function $B_1$ given in the paper. These relations carry the argument: they determine every signal rate used in the fits, and they encode the UV sensitivity, since the cutoff $\\Lambda$ appears logarithmically. The paper's sensitivity to UV origin comes from the fact that different UV completions map to different points in the bare-coupling plane, so the shape of the exclusion contour can in principle discriminate them.","core_discovery":"The central claim is that the LHC already constrains ALPs whose tree-level interactions are confined to the top quark and gluons, because loop corrections generate the couplings that make the ALP observable. Concretely, a bare coupling $c_t^0$ to tops induces an effective ALP–gluon coupling $c^\\mathrm{eff}_{agg} = c_{\\tilde G}^0 - (\\alpha_s/8\\pi) c_t^0 B_1(4m_t^2/p^2)$, while a bare gluon coupling $c_{\\tilde G}^0$ induces an effective ALP–top coupling $c^\\mathrm{eff}_t \\simeq c_t^0 + 0.6\\,c_{\\tilde G}^0$ at typical LHC energies. These loop-induced couplings, combined with the assumption that the ALP is invisible at the detector, turn measurements of the transverse momentum of tops, jets, photons, and Z bosons into exclusion limits. The resulting 2σ bounds are $f_a/c_t^0 \\gtrsim 200$ GeV and $f_a/c_{\\tilde G}^0 \\gtrsim 5$ TeV, and they are sensitive to whether the UV theory has a nonzero $c_t^0$ or $c_{\\tilde G}^0$ at the matching scale.","pith_inferences":["Extending the same loop-aware reinterpretation to ALP couplings to bottom quarks or tau leptons would likely produce complementary bounds, since the loop functions and mixing structure follow the same pattern.","If the ALP is ever discovered, the shape of the exclusion contour in the $(c_t^0, c_{\\tilde G}^0)$ plane could act as a fingerprint of the UV completion—for instance, distinguishing a pure anomaly-induced gluon coupling from one generated by top or top-partner loops, since the sign and energy dependence of the loop corrections differ.","The invisible-ALP assumption, while motivated by dark-sector scenarios, is conservative for the missing-energy channels; adding visible-decay signatures (e.g., displaced vertices or diphoton resonances) would likely close additional regions of parameter space.","Combining these LHC bounds with astrophysical and low-energy constraints on $f_a$ could further narrow the allowed parameter space for composite-Higgs-like UV models, and may motivate dedicated searches in the semileptonic top-pair channel at the HL-LHC."],"forward_implications":["Run 3 and HL-LHC data should sharpen the constraints: the expected 2σ exclusion contours extend further in the same plane under simple luminosity scaling.","The semileptonic top-pair final state ($\\ell + E_T^\\mathrm{miss} +$ jets) is identified as the single most sensitive channel for $|c_t^0|$, so future differential top-pair measurements will directly improve the ALP–top bound.","Because loop mixing generates one coupling from the other, a UV theory that starts with only $c_t^0$ or only $c_{\\tilde G}^0$ at the matching scale still produces observable signals in both top-pair and dijet channels; no hidden 'top-only' or 'gluon-only' ALP model escapes these constraints.","The domain-of-validity check shows the most sensitive signal regions have $\\sqrt{\\hat s}$ of a few hundred GeV, safely below $f_a$, so the bounds are not artifacts of pushing the EFT beyond its regime."],"supporting_citations":[{"why":"Supplies the one-loop effective ALP couplings and loop functions (Eqs. 2.11–2.25) that form the core of the analysis.","marker":"[46]"},{"why":"Provides the EFT Lagrangian and UFO model used to simulate ALP production at the LHC.","marker":"[9]"},{"why":"Established the collider-stability and EFT-validity framework for ALP–top couplings that this paper extends.","marker":"[12]"},{"why":"Defines the global reinterpretation method that combines many LHC measurements into likelihood-based exclusions.","marker":"[37]"},{"why":"Provides the CMS semileptonic top-pair differential data that dominate the $|c_t^0|$ exclusion.","marker":"[52]"},{"why":"Provides the ATLAS inclusive jet and dijet differential data that dominate the $|c_{\\tilde G}^0|$ exclusion.","marker":"[78]"},{"why":"Provides ATLAS dilepton top-pair differential data used in the fully leptonic channel.","marker":"[74]"},{"why":"Provides ATLAS all-hadronic top-pair differential data used in the fully hadronic channel.","marker":"[67]"}],"fun_headline_variants":["Loop corrections make axion–top–gluon couplings visible at LHC","Existing LHC measurements probe loop-induced axion effects","Axion–top–gluon plane squeezed by loop-generated signals","LHC data reinterpreted to bound axions with top and gluon ties","No new collider needed: Run 2 already probes axion–top couplings"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's limits rely on the assumption that ALPs produced in the final state, and whatever they decay into, escape the detector without being seen, so the signal is missing energy; if the ALP instead decays visibly, the constraints shift, although a version using only virtual ALP exchange still leaves bounds in place.","fun_headline_variants_meta":{"raw":{"variants":["Loop corrections make axion–top–gluon couplings visible at LHC","Existing LHC measurements probe loop-induced axion effects","Axion–top–gluon plane squeezed by loop-generated signals","LHC data reinterpreted to bound axions with top and gluon ties","No new collider needed: Run 2 already probes axion–top couplings"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000942,"raw_usage":{"total_tokens":4097,"prompt_tokens":1088,"completion_tokens":3009,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":704,"completion_tokens_details":{"reasoning_tokens":2912}},"tokens_in":704,"tokens_out":3009,"duration_ms":21748,"temperature":1.0,"reasoning_tokens":2912,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:38:52.563893+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An independent calculation of the one-loop mixing coefficient that produces $c^\\mathrm{eff}_t$ from $c_{\\tilde G}^0$ (the factor multiplying $c_{\\tilde G}^0$ in Eq. (2.22), shown to be about 0.4–0.7 in the relevant energy range) that disagrees with the paper's result would invalidate the mapping between bare and effective couplings and therefore the quoted exclusion limits.","supporting_citations":[],"review_version":2}