{"id":"d7937520-2374-40da-bd59-881bfee40cbd","arxiv_id":"2608.10064","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"SMEFT electroweak precision predictions can be upgraded to NLL accuracy using two-loop renormalization-group evolution plus one-loop matching, yielding complete 210-coefficient expressions and stronger indirect new-physics constraints.","lead":"This paper derives next-to-leading-logarithmic predictions for electroweak precision observables in the Standard Model Effective Field Theory, covering all 210 independent Wilson coefficients. The upgrade matters because future e+e- colliders like FCC-ee and CEPC will measure these observables precisely enough that old leading-logarithm formulas are insufficient.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"RGE solution truncated at L^2: higher-order LL/NLL logarithms from one-loop running are omitted, so the claimed NLL accuracy in Eqs. (2.8)-(2.20) and §4 is not established by the paper's own counting.","rationale":"The reader's weakest_assumption focused on the veracity of the imported two-loop anomalous dimension matrix [20]. That is a legitimate external dependency, but it is not the most load-bearing issue: even if [20] is correct, the paper's construction of the NLL predictions is internally incomplete relative to its own logarithmic counting. The paper defines LL and NLL as infinite series in Eqs. (1.2)-(1.3), yet the displayed results contain only the lowest-order term of each class (L^2 and L for the Q_uu example). The one-loop RGE used for the LL part is not solved to all orders; diagonal entries such as γ_{Hu,Hu} generate α_t^3 L^3 and higher LL terms, and the two-loop beta functions combined with one-loop iterations generate α_t^3 L^2 NLL terms. These are omitted. The impact is numerically relevant at the stated precision (α_t L ~ 0.18). This is an internal consistency concern, not a disagreement with consensus: the formal definitions in Section 1 and the implemented truncation in Sections 2-4 are mismatched. The concrete test—solving the one-loop RGE exactly for a simple subsystem—is straightforward and would settle whether the omitted terms are actually present. If present, the central claim of systematic NLL improvement needs qualification or the calculation must be extended to full one-loop exponentiation. The paper otherwise contains valuable results (classification of two-loop operator effects, model applications), so the appropriate disposition is conditional acceptance rather than rejection.","tokens_in":53157,"tokens_out":30774,"duration_ms":286949,"concrete_test":"Take the minimal subsystem {Q_uu, Q_Hu, Q_HD} (and, if needed, Q_qu) with the one-loop anomalous dimensions from Refs. [17-19]. Solve the RGE exactly (e.g., by numerical matrix exponentiation, or analytically for the subsystem) to obtain C_HD(mZ)/C_uu(Λ) with input couplings from Eq. (4.3) and Λ=1 TeV. Compare the exact one-loop result, which contains α^n L^n terms for all n, with Eq. (2.8). If the exact result differs from Eq. (2.8) by more than 10% of the L^2 coefficient, or if a nonzero α_t^3 L^3 term is present, then Eq. (2.8) is not the LL-resummed expression and the NLL numerical predictions in Section 4 omit formally required contributions. A positive outcome of this check would require the authors to either resum the one-loop RGE or revise the NLL claim.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper defines LL as α^n L^n and NLL as α^n L^{n-1} in Eqs. (1.2)-(1.3), with n=1,2,... . The central results, e.g. Eq. (2.8) for C_HD(mZ)/C_uu(Λ) and the Section 4 numerical formulas, contain only L^2 and L terms. However, solving the one-loop SMEFT RGE (which is used for the LL part) for the chain Q_uu -> Q_Hu -> Q_HD gives, in addition to the α_t^2 L^2 term, a term γ_{HD,Hu} γ_{Hu,Hu} γ_{Hu,uu} L^3/6 + ... if the one-loop diagonal anomalous dimension γ_{Hu,Hu} of Q_Hu is nonzero. Since γ_{Hu,Hu} is generically nonzero (the operator Q_Hu is not conserved; it receives contributions from g_Y, g_L, y_t), this α_t^3 L^3 term is part of the LL series (n=3) and the corresponding α_t^3 L^2 terms are part of the NLL series (n=3) by the paper's own definitions. The paper never exhibits or accounts for these higher-order logarithms, and in Section 3 the RGE solution for C_Hu is explicitly truncated at O(L), see Eq. (3.3). Thus the expressions in Sections 2-4 are fixed-order truncations at two loops/order L^2, not NLL-resummed results. The numerical impact is not negligible: for Λ=1 TeV, α_t L ~ 0.18, so the omitted α_t^3 L^3 term is expected to be roughly 10-20% of the α_t^2 L^2 term. This threatens the central claim that the predictions are systematically improved to NLL accuracy and that the fits in Section 5 (especially the projected FCC-ee sensitivities in Section 6.3) reflect true NLL precision.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims to systematically improve SMEFT predictions for electroweak precision observables (EWPO) to next-to-leading-logarithmic (NLL) accuracy by combining tree-level and one-loop matching with one- and two-loop renormalization-group evolution. It classifies operators according to the loop order at which they first enter EWPO, derives two-loop RG-induced logarithmic contributions for representative operators such as the right-handed four-top operator, the top chromomagnetic dipole, the triple-gluon operator, the top Yukawa-type operator, and the Higgs-gluon operator, and provides numerical NLL formulas for the complete set of EWPO in the LEP scheme. These results are then used to perform fits to current LEP/SLC data and projected FCC-ee data, with comparisons to LHC top-quark and Higgs constraints, and are applied to custodial Randall-Sundrum and composite Higgs models. The central claim is that the paper provides NLL-accurate EWPO expressions depending on all 210 independent Wilson coefficients, with ancillary files containing the full set of formulas.","tokens_in":53599,"tokens_out":5520,"duration_ms":54946,"significance":"If the NLL claim is correct, this work would be a valuable step for global SMEFT fits, since two-loop RG effects would allow EWPO to probe several operator classes that are otherwise only weakly constrained. The paper has clear strengths: it engages with the recently computed two-loop SMEFT anomalous dimension matrix, illustrates the mechanism on explicit operator-observable pairs, provides reproducible numerical inputs, includes comparisons with existing LHC constraints, and gives explicit model applications. The analytic and numerical exercises in Sections 2.3 and 4 appear internally consistent with the cited one-loop matching and two-loop beta-function results. However, the central NLL accuracy claim is not supported by the paper's own logarithmic counting, as detailed in the major comments. The significance of the phenomenological results is therefore conditional on a re-scoped or corrected treatment of higher-order logarithms.","major_comments":[{"comment":"The paper's own definitions in Eqs. (1.2)–(1.3) count LL as α^n L^n and NLL as α^n L^{n-1} for all n=1,2,..., yet the central results in Eqs. (2.8)–(2.20) and all of Section 4 contain only L^2 and L terms. Under the same counting, the one-loop RGE that generates the LL series also produces higher-order terms: for the chain Q_uu → Q_Hu → Q_HD, a non-zero one-loop diagonal anomalous dimension γ_{Hu,Hu} gives a contribution γ_{HD,Hu} γ_{Hu,Hu} γ_{Hu,uu} L^3/6 to C_HD(mZ), which is an n=3 LL term according to Eq. (1.2), together with corresponding n=3 NLL terms. Since α_t L ≈ 0.18 for Λ=1 TeV, the omitted α_t^3 L^3 term is expected to be roughly 10–20% of the retained α_t^2 L^2 term. The expressions are therefore fixed-order truncations at two-loop/order-L^2, not NLL-resummed results. The central claim of systematic NLL accuracy is not established without resumming or otherwise systematically including the all-order one-loop logarithms.","section":"Section 2.3, Eqs. (2.8)–(2.20)"},{"comment":"The argument that one-loop matching suffices for NLL accuracy is incomplete for the same reason. In Eq. (3.3), the RGE solution for C_Hu(μR) is explicitly truncated at O(L). If the one-loop self-mixing of Q_Hu is retained, C_Hu(μR) receives α_t^2 L^2 and higher terms, which through the one-loop matching contribution T^(1) in Eq. (3.2) produce α_t^3 L^3 and α_t^3 L^2 contributions. These are part of the LL and NLL series by the paper's own definitions and are not contained in T_NLL defined in Eq. (3.5). The statement that the remainder T−T_NLL contains no large logarithms only isolates the two-loop matching term T^(2); it does not account for omitted higher-order RG logarithms generated by the one-loop anomalous dimensions.","section":"Section 3, Eq. (3.3) and Eqs. (3.5)–(3.6)"},{"comment":"Because all numerical formulas in Section 4 contain only L^2 and L terms, the fits in Section 5, and especially the FCC-ee projections, are based on the truncated O(L^2) expressions. The projected effective scales in Table 6, such as 20.9 TeV for c_{qq}^{(1)3333}, therefore inherit the missing higher-order logarithms. The estimated numerical impact from the omitted terms is at the 10–20% level for Λ=1 TeV and grows with Λ. The quantitative sensitivities presented in Tables 5–6 and Figures 8–11 should be recomputed with a consistent treatment of higher-order logarithms, or the claims should be re-scoped to 'two-loop RG-improved predictions up to order L^2' rather than 'NLL accuracy'.","section":"Section 4 and Section 5, Tables 5–6 and Figures 10–11"}],"minor_comments":[{"comment":"The sentence describing 'the L^2, L, and L dependence' appears to contain a typo or an undefined notation; please define separate symbols for ln(Λ/mZ), ln(mZ/mt), and any other logarithms appearing in the numerical formulas.","section":"Section 4, introductory paragraph"},{"comment":"The three-term notation in formulas such as Eq. (4.5) is confusing because all three terms are written with the same symbol L; please clarify which terms are L^2, which are L, and which are constants or other logarithms.","section":"Section 4, Eqs. (4.5)–(4.22)"},{"comment":"The column header 'Future precision 10^-4' is ambiguous for mW and ΓW, whose entries appear to have dimensionful units; please clarify the units and normalization used in that column.","section":"Table 4"},{"comment":"The phrase 'the instant classic [20]' is colloquial and not appropriate for a journal report; please rephrase.","section":"Section 1, footnote 1"},{"comment":"The statement that 'LL and NLL contributions are highlighted in red and green' will not be visible in black-and-white printing; please add explicit labels or a distinct notation.","section":"Section 2.3, Eq. (2.8)"}],"recommendation":"major_revision","confidential_remarks":"The paper contains useful and probably correct fixed-order O(L^2) results, and the phenomenological analysis is careful, but the central NLL accuracy claim is not supported by the paper's own counting. The authors should either resum the one-loop RGE series to all orders (or at least include the higher-order terms systematically), or re-scope the terminology. This is a fixable issue within the manuscript's scope, so major revision rather than rejection seems appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing: this is a solid, useful paper, but the NLL label doesn't hold up in the way the abstract claims. What’s new and good: systematic classification of all 210 operators by EWPO sensitivity, explicit two-loop mixing examples, full numerical expressions in ancillary files, and a clean demonstration that one-loop matching suffices for the logarithms that matter. The LEP/SLC fits and the RS/CH examples are carefully done, with honest treatment of uncertainties. The paper deserves credit for making the two-loop beta functions of [20] usable in EWPO phenomenology.\n\nThe stress-test concern is real. By the paper's own definitions, LL means α^n L^n and NLL means α^n L^{n-1} for all n≥1. The central results (2.8)–(2.20) and the Section 4 formulas contain only L^2 and L terms. Solving the one-loop RGE along the Q_uu → Q_Hu → Q_HD chain generates α_t^3 L^3 and higher LL logarithms because γ_{Hu,Hu} is nonzero. Those terms are not exhibited or included. The paper is therefore a fixed-order two-loop calculation, not an NLL-resummed one. The numerical size is not negligible: α_t L ~ 0.18 at Λ = 1 TeV, so the missing α_t^3 L^3 term is an O(10–20%) correction to the α_t^2 L^2 term. For FCC-ee projections, that can shift the bounds by more than the stated experimental precision.\n\nAlso, the paper relies completely on the two-loop anomalous dimension matrix of [20] without independent check. That is a legitimate external benchmark, not a defect, but it means the central numbers inherit any errors in [20].\n\nSoft spots in proportion: the classification and formulas are reproducible; the fits are documented. The main revision needed is to reframe the claims honestly — \"log-enhanced corrections through two-loop order\" rather than \"NLL accuracy\" — and possibly add the missing higher-log terms or estimate their size. Some of the conclusions in Sections 5 and 6 may change once those terms are included.\n\nBottom line: this is a paper for SMEFT phenomenologists, and it deserves a serious referee. If the authors either rename the accuracy or (better) include the higher-order logarithms, it will be a valuable reference. I would read it again, and I'd cite the operator classification.\n\nRecommendation: send to peer review, with a request for clarification of the logarithmic counting.","headline":"A useful fixed-order two-loop EWPO package, but the 'NLL' label overclaims: the logarithms are truncated at L^2 and L, not resummed.","tokens_in":54106,"tokens_out":3129,"would_cite":true,"duration_ms":29822,"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":"Two-loop running upgrades electroweak precision predictions in the SMEFT to next-to-leading-logarithmic accuracy, and at that accuracy Z-pole data already set the best bounds on key four-top operators.","keywords":["SMEFT","electroweak precision observables","next-to-leading logarithms","two-loop anomalous dimensions","Wilson coefficients","four-top operators","FCC-ee","composite Higgs"],"falsifier":"Take one specific two-loop mixing entry used in the paper—for example the coefficient of the $L$ term in $C_{HD}(m_Z)/C_{uG}^{33}(\\Lambda)$ in Eq. (2.11), which is proportional to $g_s y_t\\,(48\\alpha_t - \\tfrac{32}{3}\\alpha_Y)/(16\\pi^3)$—and compute it independently with a second two-loop anomalous-dimension calculation. If the coefficient differs, the NLL shifts in $m_W$, $\\Gamma_Z$, and the asymmetries induced by $Q_{uG}$ change accordingly. Alternatively, compute the full two-loop matching contribution of $Q_{uu}^{3333}$ to $T$ and check whether it is free of $\\ln(\\Lambda/m_Z)$ terms, as the paper's claim that two-loop matching is NNLL requires.","tokens_in":52978,"feed_emoji":"🔬","tokens_out":10674,"duration_ms":97527,"temperature":0.7,"pith_summary":"Electroweak precision observables—the Z and W masses and widths, cross sections, and asymmetries measured at LEP and SLC—are among the sharpest indirect tests of new physics, but their SMEFT predictions have until now been limited to leading-logarithmic and fixed-order accuracy. This paper shows that next-to-leading-logarithmic (NLL) accuracy is already achievable: combine the one-loop matching results that exist for every electroweak observable with the recently completed two-loop anomalous dimension matrix for the full set of dimension-six operators, and evolve from the new-physics scale to the Z pole. The large logarithms of the scale ratio are generated only by this running, so two-loop matching—still unavailable for most operators—is not needed. The resulting NLL formulas cover all 210 independent Wilson coefficients and, when fitted to current data, already make LEP and SLC the strongest probes of the purely right-handed four-top operator and several related third-generation four-quark operators, with projected FCC-ee sensitivity reaching effective scales of order 20 TeV.","feed_headline":"Two-loop running sharpens electroweak precision probes of new physics","feed_subtitle":"LEP and SLC data already beat LHC limits on several four-top operators; FCC-ee would push reach to 20 TeV.","key_machinery":"The load-bearing object is the two-loop anomalous dimension matrix of the dimension-six SMEFT operators—the $210\\times 210$ matrix that describes how Wilson coefficients mix as the renormalization scale slides from $\\Lambda$ down to $m_Z$. The method combines this with the complete one-loop electroweak matching conditions, in the LEP input scheme $(\\alpha, G_F, m_Z)$. The step that makes the whole construction work is the scale-separation argument: logarithms of $\\Lambda/m_Z$ can only arise from RG running, never from loops at the electroweak scale, so one-loop matching plus two-loop running is exactly what NLL accuracy requires. The paper also uses a consistent classification of operators by the order at which they first enter—tree, one-loop, or two-loop—with the two-loop class subdivided into operators that get LL contributions through RG mixing and those (like $Q_G$, $Q_{uH}$, $Q_{HG}$) whose EWPO sensitivity first appears as genuine two-loop NLL effects.","core_discovery":"The central claim, stated on the authors' terms, is that the logarithmic structure of electroweak precision predictions in the SMEFT can be systematically improved one step beyond leading logarithms. Writing the large logarithm as $L=\\ln(\\Lambda/m_Z)$, LL terms are $\\alpha^n L^n$ and NLL terms are $\\alpha^n L^{n-1}$; the paper derives NLL predictions for all electroweak observables by feeding two-loop RG evolution into the known one-loop matching conditions. A key part of the claim is a structural argument, illustrated on $Q^{3333}_{uu}\\to T$, that two-loop matching does not generate $\\ln(\\Lambda/m_Z)$ terms and therefore enters only at NNLL order. On this basis the paper classifies every dimension-six operator by the loop order at which it first affects EWPO, identifies the Higgs self-interaction operator $Q_H$ as the unique operator whose leading logarithms appear only at four-loop order, and provides numerical fits showing that existing Z-pole data already surpass LHC top-quark constraints on the third-generation four-quark operators, with FCC-ee projections extending the reach.","pith_inferences":["The strong $\\Lambda$ dependence of the EWPO bounds on four-quark operators means that a global fit treating $\\Lambda$ as a free parameter—rather than fixing it at 1 TeV—could change the ranking of constraints; the paper fixes $\\Lambda$ but notes that larger $\\Lambda$ strengthens the RG effects, leaving this a direct, testable next step.","If the same two-loop anomalous dimension matrix is applied to dimension-eight SMEFT operators when it becomes available, the same NLL counting would predict new first-access operators for EWPO; the classification logic in Section 2.3 gives a template for identifying them.","The paper's identification of $m_W$ as the dominant NLL probe for several operators suggests that the planned FCC-ee $m_W$ measurement is not just a Standard Model consistency check but a high-sensitivity search for top-quark compositeness; experiments might therefore want to optimize $m_W$ systematics beyond the S1 scenario assumptions."],"forward_implications":["Existing LEP and SLC electroweak data already provide stronger 95% CL bounds on the third-generation four-quark operators $Q_{qq}^{3333(1)}$, $Q_{qq}^{3333(3)}$, $Q_{qu}^{3333(1)}$, and $Q_{uu}^{3333}$ than current LHC top-quark measurements, with effective scales above 1 TeV.","Projected FCC-ee Tera-Z measurements extend these bounds to effective scales of order 20 TeV and surpass projected HL-LHC sensitivity for $Q_G$, $Q_{H\\square}$, $Q_{uG}^{33}$, and $Q_{uH}^{33}$.","Operators such as $Q_{uG}^{33}$, $Q_G$, $Q_{uH}^{33}$, and $Q_{HG}$ first become accessible to electroweak precision through genuine two-loop NLL effects, while $Q_H$ remains beyond NLL reach.","NLL accuracy requires no two-loop matching; every NLO electroweak calculation whose one-loop matching is known can be upgraded to NLL by adding the two-loop RG running.","The W-boson mass is the single most important EWPO for these NLL constraints in several operator classes, dominating fits for $Q_{uu}$, $Q_{uG}$, $Q_{uH}$, and $Q_{H\\square}$."],"supporting_citations":[{"why":"Supplies the two-loop anomalous dimension matrix of all dimension-six SMEFT operators, the central input for NLL RG evolution.","marker":"[20]"},{"why":"Provides the complete NLO fixed-order EWPO predictions in five input schemes, used as the tree-level and one-loop matching conditions.","marker":"[11]"},{"why":"Provides the two-loop matching contribution to the T parameter used to demonstrate that two-loop matching is not needed for NLL accuracy.","marker":"[15]"},{"why":"Supplies the one-loop SMEFT beta functions from which the LL evolution and part of the NLL structure are derived.","marker":"[17–19]"},{"why":"Provides the LEP and SLC experimental measurements of electroweak precision observables that anchor the current-data fits.","marker":"[1]"},{"why":"Concurrent NLL EWPO analysis that this work extends to the complete set of relevant operators and to the full set of EWPO.","marker":"[45]"},{"why":"Supplies the FCC-ee Tera-Z uncertainty projections (S1 scenario) used for the future-sensitivity fits.","marker":"[186]"},{"why":"Previous fixed-order NNLO analysis of third-generation four-quark operators in EWPO to which the NLL results are compared.","marker":"[16]"}],"fun_headline_variants":["NLL logs sharpen SMEFT fits to electroweak data","Two-loop RG boosts precision of SMEFT predictions","Z-pole data beat LHC on four-top operators","FCC-ee could probe new physics to 20 TeV via EWPO","All 210 Wilson coefficients now at NLL in EWPO"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper takes the two-loop scale-evolution matrix for all 210 operators from a separate calculation instead of deriving it, and it only demonstrates in one worked example that one-loop matching is enough; if that matrix has an error, or if the worked example does not represent every operator, every NLL number and bound in the paper shifts.","fun_headline_variants_meta":{"raw":{"variants":["NLL logs sharpen SMEFT fits to electroweak data","Two-loop RG boosts precision of SMEFT predictions","Z-pole data beat LHC on four-top operators","FCC-ee could probe new physics to 20 TeV via EWPO","All 210 Wilson coefficients now at NLL in EWPO"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000437,"raw_usage":{"total_tokens":2220,"prompt_tokens":945,"completion_tokens":1275,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":561,"completion_tokens_details":{"reasoning_tokens":1190}},"tokens_in":561,"tokens_out":1275,"duration_ms":8031,"temperature":1.0,"reasoning_tokens":1190,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:14:49.861307+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take one specific two-loop mixing entry used in the paper—for example the coefficient of the $L$ term in $C_{HD}(m_Z)/C_{uG}^{33}(\\Lambda)$ in Eq. (2.11), which is proportional to $g_s y_t\\,(48\\alpha_t - \\tfrac{32}{3}\\alpha_Y)/(16\\pi^3)$—and compute it independently with a second two-loop anomalous-dimension calculation. If the coefficient differs, the NLL shifts in $m_W$, $\\Gamma_Z$, and the asymmetries induced by $Q_{uG}$ change accordingly. Alternatively, compute the full two-loop matching contribution of $Q_{uu}^{3333}$ to $T$ and check whether it is free of $\\ln(\\Lambda/m_Z)$ terms, as the paper's claim that two-loop matching is NNLL requires.","supporting_citations":[],"review_version":1}