{"id":"106a0e38-3b16-4c7f-85df-a5a7e93416a9","arxiv_id":"2511.02910","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"With renormalisation-group running included, the one-loop high-temperature Daisy-resummed potential at µ=πT reproduces the phase-transition parameters of the two-loop dimensionally reduced EFT, while the no-running one-parameter approximation does not.","lead":"This paper compares three ways of calculating the strength of a supercooled dark-sector phase transition that could explain the nano-Hertz gravitational-wave background seen by pulsar timing arrays, and shows that adding renormalisation-group running makes the simple one-loop method agree with the two-loop effective-field-theory benchmark. The result matters because it tells theorists which shortcuts are reliable when predicting gravitational-wave signals from early-universe","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"3D 'reduced scale dependence' is demonstrated by varying only µ_Match, not the soft scale µ3; the comparison with 4D HT may be biased and the central claim not yet established.","rationale":"I focused on the most load-bearing support for the central comparative claim: the demonstration that the 3D EFT has reduced residual scale dependence. The reader's weakest assumption was the 3D power-counting validity in the supercooled regime (a real, self-acknowledged limitation). However, that concern is qualitative: formally invalid power counting does not automatically make the two-loop result numerically unreliable, and the paper already labels the 3D two-loop result as 'best available approximation'. A sharper, falsifiable gap is that the scale-dependence comparison is asymmetric: the 3D band in Figs. 2 and 6 varies only µ_Match, while the 3D EFT has two scales (µ_Match and µ3). Since the 3D potential depends explicitly on µ3 via Eq. (A16), holding µ3 fixed may underestimate the true residual scale dependence. If varying µ3 widens the band to the level of the 4D HT band, the paper's headline 'significantly reduced scale dependence' loses its quantitative basis. This does not require challenging the benchmark accuracy, and it is directly testable with the existing DRalgo tool. Because the concern is conditional — the claim may survive the test — the verdict remains CONDITIONAL as the reader proposed; no change to the verdict is needed, but the condition should include this check.","tokens_in":24836,"tokens_out":11494,"duration_ms":120295,"concrete_test":"Recompute the 3D two-loop (NLO) phase-transition parameters with DRalgo for g=0.5 and g=0.8, varying both µ_Match and µ3 independently over the ranges µ_Match∈[2πT/4, 4πT] and µ3∈[gT/4, 4gT] (e.g., a 3×3 grid). For each combination, compute S3/T, Tn, β/H, α and the GW peak amplitude. If the envelope over both scales is comparable in width to the 4D HT band at µ∈[πT/4, 4πT], the 'significantly reduced scale dependence' claim is not established; if it remains much narrower, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim that the 3D EFT gives a substantially reduced renormalisation-scale dependence is supported by the purple bands in Fig. 2 (right) and Fig. 6 (right), which quantify the residual µ-dependence of the 3D two-loop (NLO) scheme. These bands are obtained by varying only the matching scale µ_Match around its reference value 2πT; the soft scale µ3 is held fixed at gT (Table I). The 3D two-loop effective potential explicitly depends on µ3 through log(µ3/µ) in Eq. (A16) and through the 3D loop logarithms. Thus the displayed 3D band covers only one of the two independent scales of the 3D EFT. In contrast, the 4D HT band (green) varies its only scale µ around πT. Comparing the widths of a one-dimensional variation (4D) with a partial two-dimensional variation (3D) biases the comparison in favour of the 3D result. If µ3 were varied as well, the 3D band could widen substantially, weakening the headline that 3D EFT has significantly reduced scale dependence. This is distinct from, but complementary to, the acknowledged power-counting limitation (Section II.c), and it directly targets the quantitative basis of the paper's benchmark recommendation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies supercooled first-order phase transitions in a classically scale-invariant dark Abelian Higgs model, comparing three theoretical treatments of the finite-temperature effective potential: the one-parameter approximation (OPA), the 4D high-temperature one-loop potential with Daisy resummation, and a dimensionally reduced 3D effective field theory computed with DRalgo at one and two loops. The authors argue that consistent RG running is essential, that the OPA without running deviates significantly, that the 3D two-loop EFT has substantially reduced renormalisation-scale dependence, and that, once the 4D HT scale is chosen as µ=πT, the 4D HT result agrees with the 3D two-loop benchmark. The bounce action is computed semi-analytically and validated against shooting, and the phase-transition parameters (Tn, α, β/H) are propagated to gravitational-wave spectra and compared with NANOGrav 15-year data.","tokens_in":25112,"tokens_out":4305,"duration_ms":48783,"significance":"If the conclusions hold, the paper provides a practically useful simplification: a one-loop 4D Daisy-resummed potential with RG running and µ=πT may be sufficient for supercooled conformal dark sectors, and the more expensive 3D two-loop DR calculation can be reserved as a benchmark. The use of DRalgo makes the 3D matching a parameter-free application of a public automated framework, and the explicit comparison of four 3D schemes is a useful systematic study. The paper is also candid about several limitations, including the validity of the NLO power counting in the deep supercooled regime. However, the central claim of reduced scale dependence in the 3D EFT is not fully established because only one of the two independent scales of the 3D EFT is varied, and the agreement between 4D HT and 3D is partly a calibration rather than an independent prediction.","major_comments":[{"comment":"The reduced scale dependence of the 3D EFT is demonstrated by varying only the matching scale µ_Match around 2πT, while the soft scale µ3 is fixed at gT (Table I). However, Eq. (A16) explicitly depends on log(µ3/µ), and the 3D loop potential itself depends on µ3. Thus the purple bands in Figs. 2 and 6 cover only one of the two independent renormalisation scales of the 3D EFT. The conclusions even state that the 3D 1-L (NLO) scheme 'introduces additional scale dependence through the matching of the soft scale µ3', but this dependence is never quantified. To support the headline claim, the authors should repeat the variation with µ3 varied (e.g., µ3 ∈ [gT/4, 4gT]) and propagate this to S3/T, Tn, β/H and the GW spectrum. If the 3D band widens substantially, the conclusion should be moderated to 'reduced matching-scale dependence' rather than 'reduced scale dependence'.","section":"§II.c, Eq. (A16), Fig. 2 (right), Fig. 6 (right)"},{"comment":"The paper itself notes that the NLO power counting assumes λ∼g^2 and therefore 'is, strictly speaking, not valid in the supercooled regime where λ≲g^4'. The benchmark model has λ(µ0)=0 and negative running, so the 3D two-loop result is used precisely in the regime where its systematic expansion may break down. Since the 4D HT scale µ=πT is then calibrated against this same 3D benchmark, any uncontrolled truncation error in the 3D potential propagates directly into the central consistency claim. A quantitative estimate of this truncation uncertainty — for example by studying the size of the available higher-order terms, by comparing LO/NLO/mixed schemes more carefully, or by varying the power-counting assumptions — is needed before the 3D result can serve as the benchmark that anchors the µ=πT choice.","section":"§II.c"},{"comment":"The choice µ=πT is not an independent prediction but is fixed by requiring the 4D HT bounce action to agree with the 3D two-loop result: 'we take the 3D two-loop NLO potential as a benchmark and fix the 4D HT renormalisation scale by requiring the good agreement with the bounce action'. Therefore the agreement between 4D HT and 3D two-loop is partly by construction. The abstract's phrase 'with a suitable choice of RGE scale' is accurate, but the stronger statement that the 4D HT calculation 'yields results consistent with the 3D calculations' should be framed as a calibration statement. An independent validation of µ=πT — for example by a principal-of-minimal-sensitivity condition on the 4D potential itself, or by showing that the calibrated scale remains stable when µ3 is varied — would substantially strengthen the argument.","section":"II (last paragraph), III.A"}],"minor_comments":[{"comment":"The captions should explicitly state that for the 3D bands the soft scale µ3 is held fixed at gT and only µ_Match is varied. Currently this information is only in Table I, and the caption wording 'varied analogously' could mislead readers into thinking the full scale variation of the 3D EFT is shown.","section":"Fig. 2 and Fig. 6 captions"},{"comment":"It would be useful to state explicitly that the beta functions and anomalous dimensions are at one-loop order, and to clarify that the RGEs are solved in the 4D MS scheme with the same reference scale µ0 used for the input parameters. This would make the matching between the 4D running and the 3D matching scale µ_Match easier to follow.","section":"Appendix B, Eq. (B1)"},{"comment":"The text says that at larger couplings 'both approximations show comparably sized deviations relative to the two-loop curve', but the figure is on a logarithmic axis and no numerical deviation is quoted. A quantitative statement, even for one benchmark point, would make the comparison more informative.","section":"III.B, Fig. 5 (right)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope and tackles a relevant question for PTA-era phase-transition phenomenology. The main technical weakness is not fatal but needs additional numerical work: the claim of reduced scale dependence in the 3D EFT should be checked by varying µ3 as well as µ_Match, and the validity of the NLO power counting in the supercooled regime should be quantified. I would be positive after these points are addressed. No concerns about citation practice or novelty disclosure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a genuinely useful comparison of thermal potential schemes for a supercooled Abelian dark sector, and the practical message—one-loop 4D HT with Daisy at µ=πT plus RG running tracks the two-loop DR result for this benchmark—is broadly credible. But the claim that the 3D EFT dramatically reduces scale dependence is supported by a narrower test than the text suggests, and the benchmark itself has an acknowledged power-counting limitation.\n\nWhat’s new: a systematic side-by-side of 4D Daisy-resummed HT, OPA, and DRalgo-based 3D EFT at one and two loops, with explicit RG running, for a conformal U(1)_D model. The validation of the semi-analytic bounce against shooting is good practice. The observation that partial higher-order corrections (3D one-loop NLO) can be worse than a well-chosen one-loop 4D scheme is a useful caution, and the finding that 3D two-loop has narrower µ-dependence than 4D HT, within the scales they vary, is plausible and worth having on record.\n\nThe soft spots, in order of importance. First, the scale-dependence comparison is skewed: the 3D band in Figs. 2 and 6 varies only µ_Match, holding µ3 fixed at gT. But Eq. (A16) shows explicit µ3 dependence through log(µ3/µ), and the 3D potential has two renormalisation scales. Varying only one of them makes the 3D band artificially narrow relative to the 4D band, which varies its only scale. The headline “significantly reduced scale dependence” needs a two-dimensional 3D scan before it is established. Second, the 3D two-loop NLO is treated as the accuracy benchmark even though Section II.c admits the power-counting assumes λ∼g^2 and is not valid for λ≲g^4, which is exactly the supercooled regime. Since µ=πT is chosen by matching this benchmark, some circularity remains; this is not a blind prediction. Third, the abstract says OPA deviates significantly, but that applies only to the no-running OPA. Figure 7 shows OPA with running largely agrees. The body is honest about this, but the abstract overstates. Finally, no code or data is shipped and there is no attempt to quantify truncation uncertainties, which is unfortunate for a paper whose message is about theoretical control.\n\nWho this is for: GW phenomenologists and model-builders who need a practical rule for which potential to use in a supercooled hidden sector. The µ=πT calibration is a useful heuristic, with the caveats above.\n\nRecommendation: send to peer review. It deserves a serious referee, and the central comparison is internally consistent and worth publishing after the scale-dependence test is expanded and the abstract qualified.","headline":"Useful, practical cross-check of 4D vs 3D potentials for supercooled dark transitions, but the reduced-scale-dependence headline rests on a one-dimensional 3D scale scan and a benchmark with an admitted power-counting caveat.","tokens_in":25727,"tokens_out":3431,"would_cite":true,"duration_ms":37975,"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":"Consistent RG running at a tuned scale makes the simple one-loop 4D high-temperature potential agree with the two-loop 3D effective theory for supercooled phase transitions, while the one-parameter shortcut fails.","keywords":["supercooled phase transitions","effective potential","Daisy resummation","dimensional reduction","renormalisation group running","bounce action","gravitational waves","conformal Abelian Higgs model"],"falsifier":"Compute the tunnelling action S3/T for the same conformal dark U(1) model directly in the 3D theory with lattice Monte Carlo at T≈0.01v and gauge coupling near the strong end of the benchmark range; if the lattice result differs from the two-loop dimensionally reduced curve by more than the residual-scale bands shown in the paper, then the 4D high-temperature agreement is calibrated to a biased benchmark.","tokens_in":24655,"feed_emoji":"🌌","tokens_out":7742,"duration_ms":77601,"temperature":0.7,"pith_summary":"This paper aims to settle how much theoretical machinery is needed to predict gravitational waves from a supercooled first-order phase transition in a classically scale-invariant dark Abelian Higgs sector. Its central claim is that a one-loop four-dimensional effective potential with high-temperature expansion and Daisy resummation, supplemented by consistent RG running of the couplings and evaluated at the renormalisation scale µ=πT, reproduces the transition parameters and gravitational-wave spectrum obtained from the more elaborate two-loop dimensionally reduced 3D effective theory. The paper also finds that the simple one-parameter approximation, when used with fixed couplings and no running, deviates strongly in the large-supercooling regime, and that the 3D EFT has markedly smaller residual renormalisation-scale dependence. The practical payoff is that a cheap and commonly used calculation can be trusted for these models once running is included, which matters for interpreting the nanohertz gravitational-wave background.","feed_headline":"RG running aligns one-loop and two-loop phase transition results","feed_subtitle":"At µ=πT the cheap one-loop calculation reproduces the two-loop 3D EFT; the one-parameter approximation misses.","key_machinery":"The load-bearing object is the finite-temperature effective potential entering the tunnelling action S3/T, the exponent that controls bubble nucleation. The paper compares three constructions: a one-loop 4D high-temperature potential with Daisy resummation, which adds thermal screening masses to the static bosonic modes; a one-parameter polynomial approximation to that potential; and a 3D effective theory obtained by dimensional reduction, in which heavy non-static modes are integrated out and the soft potential is computed to two loops. Running of the couplings via the renormalisation-group equations connects the schemes: evaluating the 4D potential at µ=πT and matching the 3D theory at 2πT","core_discovery":"Central claim: once RG running is included and the scale chosen deliberately, the standard 4D one-loop high-temperature potential with Daisy resummation closes the gap to the two-loop dimensionally reduced 3D effective theory. For a conformal dark U(1) scalar-gauge model, the authors compute the tunnelling action S3/T and extract T_n, α, β/H and the gravitational-wave spectrum. The 4D potential at µ=πT reproduces the 3D two-loop benchmark; the one-parameter approximation without running deviates sharply at large supercooling. The 3D theory has much smaller residual scale dependence; NLO inputs on a one-loop 3D potential can worsen agreement.","pith_inferences":["A testable extension: repeat the same scale-calibration exercise in a non-Abelian or multi-field conformal model; if µ=πT remains the agreement point, the recipe becomes a general shortcut for scanning supercooled transitions.","Because the 3D two-loop benchmark itself assumes λ∼g^2, which the paper notes is violated when λ≲g^4, the agreement between 4D and 3D could partly reflect shared truncation errors; a lattice computation of the 3D theory would expose this.","The close agreement between 4D and 3D at low temperature suggests the thermal barrier is controlled by the same resummed physics in both schemes; if confirmed, simpler 4D codes can be used for pulsar-timing-array interpretation without two-loop machinery.","The paper's OPA comparison is deliberately unfavourable (no running, no Daisy terms); since the appendix shows OPA with running approaches the 4D high-temperature result, the substantive lesson is not that the parameterisation is inherently bad but that running is essential."],"forward_implications":["For a classically scale-invariant dark Abelian Higgs sector, the one-loop 4D high-temperature potential with Daisy resummation and RG running at µ=πT is a reliable, cheap substitute for the two-loop 3D EFT when predicting T_n, α, β/H and gravitational-wave spectra.","Neglecting RG running, as in the one-parameter approximation used here, shifts T_n and β/H substantially at small gauge coupling, produces an unphysical peak in β/H, and gives a gravitational-wave spectrum that lies away from the RG-improved predictions.","The two-loop dimensionally reduced EFT is the most scale-stable scheme; the residual µ-dependence visible in its potential and spectra is much narrower than in the 4D high-temperature scheme.","Including NLO masses and couplings in a one-loop 3D potential can increase scale dependence; partial higher-order corrections do not automatically improve predictions.","The authors caution that the calibrated agreement is not automatically valid for other models or the non-conformal regime."],"fun_headline_variants":["RG scale choice aligns cheap and full phase-transition calculations","One-loop 4D potential matches two-loop 3D EFT at µ=πT","Daisy-resummed one-loop beats analytic bounce action in supercooling","Running couplings fix supercooled PT scale dependence","3D EFT reduces scale dependence in supercooled phase transitions"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the two-loop dimensionally reduced 3D effective theory is an accurate benchmark in the supercooled regime, even though the paper itself notes that its power counting assumes λ∼g^2 and is not valid where λ≲g^4; if that benchmark carries uncontrolled truncation errors, the calibrated scale µ=πT and the claimed 4D–3D consistency lose their anchor.","fun_headline_variants_meta":{"raw":{"variants":["RG scale choice aligns cheap and full phase-transition calculations","One-loop 4D potential matches two-loop 3D EFT at µ=πT","Daisy-resummed one-loop beats analytic bounce action in supercooling","Running couplings fix supercooled PT scale dependence","3D EFT reduces scale dependence in supercooled phase transitions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000374,"raw_usage":{"total_tokens":1846,"prompt_tokens":770,"completion_tokens":1076,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":514,"completion_tokens_details":{"reasoning_tokens":983}},"tokens_in":514,"tokens_out":1076,"duration_ms":8202,"temperature":1.0,"reasoning_tokens":983,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T00:03:23.171931+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the tunnelling action S3/T for the same conformal dark U(1) model directly in the 3D theory with lattice Monte Carlo at T≈0.01v and gauge coupling near the strong end of the benchmark range; if the lattice result differs from the two-loop dimensionally reduced curve by more than the residual-scale bands shown in the paper, then the 4D high-temperature agreement is calibrated to a biased benchmark.","supporting_citations":[],"review_version":1}