{"id":"3ade7f1e-e6c8-4a7b-8d60-e09521eafc2e","arxiv_id":"2506.21854","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"A global QCD fit to deep-inelastic scattering prefers a dark photon with mass 4 to 6 GeV and mixing 0.06 to 0.12, claimed as a 6.5 sigma hint, though direct searches exclude these parameters unless extra dark matter couplings are invoked.","lead":"A global fit to decades of electron-proton scattering data prefers adding a dark photon with 4 to 6 GeV mass and mixing of 0.06 to 0.12, claiming a 6.5 sigma preference over the Standard Model. If true this is a major discovery, but the parameters conflict with direct search limits unless the dark photon also couples to dark matter.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 6.5σ evidence claim hinges on a Δχ² improvement that could be absorbed by more flexible PDFs; the paper does not provide the statistical test or a baseline-flexibility check needed to rule this out.","rationale":"The reader identifies the PDF parametrization and missing higher-order uncertainty treatment as the weakest assumption; this stress-test agrees and makes the concern more specific. The central claim is not that dark photons exist, but that a particular global fit prefers them at 6.5σ. For that to be true, the χ² improvement must survive two checks: (i) the baseline SM fit must not be artificially inflexible, and (ii) the quoted significance must be a valid frequentist statement rather than a maximum over a parameter scan without a trials factor. This paper, as a proceedings summary, provides neither the full statistical procedure nor a flexibility control. It does however give enough information in Table 1 to see that the improvement is dataset-selective and modest in per-dof terms, which makes the PDF-absorption scenario plausible. The recommended verdict remains CONDITIONAL: the claim is checkable in principle, and the published longer analysis [21] may well contain the missing details, but the current manuscript does not yet establish robustness. If the proposed control experiment shows that a more flexible SM baseline achieves a comparable Δχ², the evidence claim would be substantially weakened; if it does not, the hint is strengthened. This is a request for a control experiment, not an accusation of error or misconduct.","tokens_in":8843,"tokens_out":5978,"duration_ms":71011,"concrete_test":"Run the SM-only JAM fit from Ref. [21] on the same 3283 data points with the same MHOU covariance, but extend the Eq. (5) parametrization by adding two or more independent shape terms (e.g. δ x^{3/2} and φ(1-x)^3) while also profiling over all PDF parameters and the dark-photon parameters M_AD, ε in a likelihood ratio test. If the baseline χ² drops by an amount comparable to the ~66 units currently attributed to the dark photon, or if the profile-likelihood significance falls below 3σ after accounting for the scan over M_AD and ε, the 6.5σ preference is not robust evidence for a dark photon.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the conversion of the Table 1 Δχ² into a 6.5σ preference. The total baseline χ²/dof is 1.05 and the dark-photon fit is 1.03 for N≈3283, i.e. Δχ²≈66; yet the improvement is concentrated in fixed-target DIS and HERA NC, while Drell-Yan, Z rapidity and jets become slightly worse (Table 1). That is the pattern expected if the dark photon parameters are acting as extra flexibility in a region where the PDF parametrization of Eq. (5) is too rigid, rather than as a coherent physical signal. The paper states that the MHOU procedure of NNPDF was used, but it gives no details of how the 6.5σ is defined (e.g. Wilks theorem, profile likelihood, scan over M_AD and ε, trials factor), and no test is shown in which the SM baseline is given comparable flexibility (more PDF parameters, or a higher-twist/target-mass term) to see whether the Δχ² persists. Since the central numerical claim of the paper is exactly this significance, this missing control is the key vulnerability.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports on what it calls the first global QCD analysis including a dark photon, claiming an improvement in chi-squared per degree of freedom from 1.05 (baseline) to 1.03 (with the dark photon) over 3283 data points, and a hypothesis-test preference for the dark photon model of 'as much as 6.5 sigma', with best-fit parameters M_AD in (4,6) GeV and epsilon in (0.06,0.12) at 95% CL. It also discusses implications for parity-violating electron scattering, rare kaon decay, and electroweak precision observables, and argues that a large dark-photon coupling to dark matter can relax direct-search constraints. The analysis is a summary of prior work by the same authors (refs [21]-[28]); the statistical test, the full fit methodology, and the uncertainty treatment are not reproduced in this manuscript.","tokens_in":9148,"tokens_out":9159,"duration_ms":88754,"significance":"If the 6.5 sigma preference were established under a rigorously defined statistical test, this would be a major indirect hint for new physics in the gauge sector. The use of a large global dataset (3283 points) with a modern uncertainty treatment (JAM framework with NNPDF-style missing higher-order uncertainties) is a strength, as is the comparison with a U(1)_B-L Z' model that does not improve the fit, which serves as a useful control. However, the central significance claim is not documented in this text, and the pattern of improvements across datasets raises concerns about PDF flexibility and a potential look-elsewhere effect. The implications sections are interesting but largely summarize the authors' earlier work rather than providing new derivations.","major_comments":[{"comment":"The central claim of a 6.5 sigma preference for the dark photon model over the Standard Model is stated without specifying the statistical test. The manuscript does not define the test statistic, the probability distribution used to convert the quoted Delta chi-squared into a significance, whether a full profile likelihood over (M_AD, epsilon) was used, or whether any trials factor for scanning the two-dimensional parameter space was applied. From Table 1, the total Delta chi-squared is approximately 66 for two new parameters; under Wilks' theorem this would correspond to about 7 sigma, while the text quotes 'as much as 6.5 sigma', suggesting a maximum over the scan and thus a need for a look-elsewhere correction. As presented, the significance cannot be reproduced, and this is the load-bearing element of the paper.","section":"Section 2 (hypothesis test paragraph)"},{"comment":"The improvement in chi-squared is concentrated in fixed-target DIS and HERA NC (together 2599 of 3283 data points), while Drell-Yan, Z rapidity and jets become slightly worse. This is the pattern expected if the two dark photon parameters are acting as extra flexibility that absorbs deficiencies of the PDF parametrization in Eq. (5) in the low-Q^2 and large-x region, rather than as a coherent physical signal. The manuscript does not provide a control fit in which the SM baseline is given comparable flexibility (e.g., additional PDF parameters, higher-twist or target-mass corrections). Without such a test, the attribution of the Delta chi-squared to a dark photon is not uniquely established.","section":"Section 2, Table 1 and Eq. (5)"},{"comment":"The best-fit region (M_AD in (4,6) GeV, epsilon in (0.06,0.12)) is above the direct search limits from BaBar [12] and CMS [14]. The manuscript argues that a coupling of the dark photon to light dark matter particles with g_chi of order unity can broaden the resonance and relax these limits by one to two orders of magnitude, citing Ref. [28]. However, g_chi is not a parameter of the DIS fit, and the relaxed constraints are not quantified in this manuscript; the compatibility of the best-fit region with the relaxed bounds is asserted qualitatively. Because the central claim concerns the existence of a dark photon, this model-dependent reinterpretation should be quantified, or explicitly flagged as a condition on the claim.","section":"Section 2, discussion following Eq. (6)"}],"minor_comments":[{"comment":"The text says 'the values of chi-squared per degree of freedom ... are given in the third column of Tab. 1', but Table 1 has separate dark and baseline columns; this should be 'the second and third columns' or 'the columns'.","section":"Section 2, near Table 1"},{"comment":"The abstract and Section 3 heading contain the typo 'party-violating electron scattering'; this should be 'parity-violating electron scattering'.","section":"Abstract and Section 3 heading"},{"comment":"The symbol M in the definition f_1(x,y) = 1 - y - x y M / (2E) is not defined; presumably it is the nucleon mass, but this should be stated explicitly.","section":"Section 3.1, Eq. (8)"},{"comment":"The covariance formula 'cov = Sigma_exp * cor * Sigma_exp' is not fully described; the correlation matrix 'cor' should be defined, and the vector notation for V and the experimental values should be made explicit.","section":"Section 3.3, Eq. (16)"},{"comment":"The notation for the dark photon mass is inconsistent: M_AD is used in the text and Eq. (6), while Figure 3 uses 'mAD' and Eq. (2) uses m_{A'}; please unify the notation.","section":"Throughout"},{"comment":"The phrase 'the first hint for the existence of a dark photon' is stronger than the evidence presented in this manuscript, which relies on a hypothesis test not fully described here; consider softening to 'an indirect hint' or 'a hint from a global QCD analysis'.","section":"Abstract and Conclusion"}],"recommendation":"major_revision","confidential_remarks":"This is a proceedings-style manuscript that heavily cites the authors' own previous papers for the central analysis. The 6.5 sigma claim is not self-contained, and the discussion of the direct-search conflict relies on an arXiv preprint [28] that is not yet fully described here. If the target venue is a conference proceedings, the expectations may be lower; however, if it is a full journal, the statistical definition and the PDF-flexibility control are essential. The 'first hint' wording is overly strong given the unresolved systematics."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis is a HADRON2025 proceedings paper summarizing the authors' own previous work. The new content is minimal: the 6.5 sigma hint is from JHEP 09 (2023) 096 [21], and the 'implications' sections are condensed from [22,23,24,27]. As a conference summary it's fine; as evidence for a dark photon it leans entirely on a significance claim that is not substantiated here.\n\nWhat the paper does well: the underlying fit is serious work by the JAM group—NLO QCD, Monte Carlo uncertainty quantification, and a missing-higher-order-uncertainty procedure taken from NNPDF. A useful control is their U(1)_{B-L} Z' analysis, which shows no chi2 improvement, indicating the effect is not generic to any new gauge boson. They also honestly flag the tension with direct detection limits and invoke a model-dependent dark-matter coupling to relax those limits.\n\nThe soft spots are the ones you'd expect from a proceedings that summarizes a big claim. The 6.5 sigma is never defined: no test statistic, no Wilks theorem or profile likelihood description, no mention of a trials factor for scanning M_AD and epsilon. Table 1 shows the improvement is concentrated in fixed-target DIS and HERA NC, while Drell-Yan, Z rapidity and jets get slightly worse—the pattern you'd expect if the dark photon parameters are effectively adding PDF flexibility. There is no baseline-flexibility check (more PDF parameters, higher-twist terms) to see if the chi2 improvement persists. They point to [21] for the full details, which is reasonable for a proceedings, but as a standalone text the central claim is not independently checkable.\n\nThe dark-matter-coupling argument for evading direct search limits is also speculative and not a falsifiable prediction here. Still, the authors are appropriately cautious in the abstract: 'first hint, albeit indirect.'\n\nOverall: worth a serious referee, but only in the context of the full analysis. I would not desk-reject the claim; a referee should demand the complete statistical treatment and a flexibility test. For a proceedings, this is acceptable but should not be cited as evidence without reading [21]. I'd bring it to a reading group only if we also assign the original JHEP paper.","headline":"A conference summary of a previously reported 6.5 sigma dark photon hint; the significance is not statistically transparent and deserves a careful look at the original fit before believing it.","tokens_in":9670,"tokens_out":4133,"would_cite":false,"duration_ms":44556,"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":"First indirect hint of a dark photon emerges from a global QCD fit.","keywords":["dark photon","global QCD analysis","deep inelastic scattering","parton distribution functions","kinetic mixing","parity-violating electron scattering","rare kaon decay","dark matter portal"],"falsifier":"A repeat of the global fit with a more flexible parton parametrization, or at next-to-next-to-leading order in the strong coupling, that shows no significant $\\chi^2$ improvement would falsify the claimed preference; a direct search sensitive to a broad resonance in the 4--6 GeV mass range that excludes the quoted $\\epsilon$ window would also do so.","tokens_in":8612,"feed_emoji":"⚛️","tokens_out":9119,"duration_ms":83720,"temperature":0.7,"pith_summary":"The paper asks whether the Standard Model's description of electron-nucleon deep-inelastic scattering and related high-energy data is improved by adding a dark photon, a new $U(1)$ gauge boson that mixes with the ordinary photon. A global QCD fit that includes dark-photon exchange is found to reduce the total $\\chi^2$ relative to the baseline fit, and a hypothesis test prefers the dark-photon model over the Standard Model by as much as $6.5\\sigma$. The paper reads this as the first indirect hint of a dark photon, with a best-fit mass between 4 and 6 GeV and kinetic mixing between 0.06 and 0.12 at 95% confidence. If the hint holds, it would open a concrete probe of new physics in the most precise scattering data available.","feed_headline":"A dark photon improves the global QCD fit by 6.5 sigma","feed_subtitle":"A new U(1) boson with mass 4-6 GeV and mixing 0.06-0.12 best fits electron-nucleon scattering data.","key_machinery":"The machinery is the global QCD fit carried out with Bayesian Monte Carlo uncertainty quantification, at next-to-leading order in the strong coupling, supplemented by the missing higher-order uncertainty procedure of [26]. Dark-photon exchange enters through the structure functions $\\widetilde F_2$ and $\\widetilde F_3$, written as sums over the physical gauge bosons $\\gamma$, $Z$, and $A'$ with weights $\\kappa_i = Q^2/(Q^2 + m_i^2)$ and the corresponding vector and axial-vector couplings; the parton distributions are parametrized at the starting scale by $f_q(x,Q_0^2)=N x^\\alpha (1-x)^\\beta (1+\\gamma\\sqrt{x}+\\eta x)$ and determined by the fit. The $\\chi^2$ difference between the dark-photon and baseline fits, combined with the number of fitted parameters, yields the hypothesis-test significance.","core_discovery":"The central claim is that a global next-to-leading-order QCD analysis of deep-inelastic scattering and related Drell-Yan, $W/Z$, and jet data is significantly improved when dark-photon exchange is added to the Standard Model, with the best dark-photon fit preferred over the baseline by as much as $6.5\\sigma$. The improvement is concentrated in the fixed-target DIS and neutral-current data, and the preferred parameters are $M_{A'}\\in(4,6)$ GeV and $\\epsilon\\in(0.06,0.12)$ at 95% CL. Because the same analysis does not favor a $U(1)_{B-L}$ $Z'$ boson, the authors argue the effect is specific to the dark photon's couplings rather than generic to any new gauge boson. They also show that the apparent tension with direct search limits can be resolved if the dark photon decays to light dark-matter particles with $\\mathcal{O}(1)$ couplings, which broadens the resonance and suppresses its detection.","pith_inferences":["The $6.5\\sigma$ significance is computed within a specific parton parametrization and uncertainty procedure; a more flexible parametrization or a full NNLO treatment could shift the significance, so the claim should be re-tested with those upgrades.","If the dark-photon interpretation is correct, high-luminosity electron-proton scattering could pin down the mass and mixing by mapping the $Q^2$-dependent distortion of the structure functions.","A dedicated search for a broad resonance in the di-muon or di-electron spectrum between 4 and 6 GeV, rather than the narrow-resonance searches performed so far, would directly confirm or exclude the best-fit region.","The $\\mathcal{O}(1)$ coupling to dark matter that relaxes direct search limits also predicts dark-matter annihilation and direct-detection signals that could be probed in dedicated experiments."],"forward_implications":["Including a dark photon in global QCD fits changes the extraction of parton distribution functions from deep-inelastic data, so derived quantities such as cross sections for future colliders should be re-evaluated with the new fit.","The dark photon produces up to 5% corrections to the weak neutral-current couplings $C_{1q}$ and $C_{3q}$ at low momentum scales and up to 10% corrections to $C_{2q}$ at $Q^2=10^3$ GeV$^2$, which will be testable in present and planned electron-scattering experiments.","In the allowed parameter region, dark-photon contributions to $\\mathrm{Br}(K_L\\to\\pi^0\\nu\\bar\\nu)$ are below 1%, so a future large anomaly in this channel would require either new physics beyond a minimally mixed dark photon or additional dark-sector couplings.","A dark photon that couples to light dark-matter particles with $\\mathcal{O}(1)$ coupling avoids current direct search bounds, making the 4--6 GeV, $\\epsilon\\sim0.06$--0.12 region a concrete target for future collider searches for broad resonances.","The comparison with a $U(1)_{B-L}$ $Z'$ boson shows that not every new gauge boson improves the global fit, strengthening the case that the signal is specifically a dark photon."],"supporting_citations":[{"why":"Supplies the theoretical treatment of dark-photon effects on proton structure functions that underlies the fit.","marker":"[18]"},{"why":"Provides the earlier deep-inelastic-scattering constraints on the dark photon that the global analysis extends.","marker":"[19]"},{"why":"Is the companion global QCD analysis with dark photons that supplies the baseline and dark-photon fits quoted here.","marker":"[21]"},{"why":"Gives the electroweak precision analysis that sets independent limits on the dark-photon coupling to dark matter.","marker":"[24]"},{"why":"Supplies the Bayesian Monte Carlo uncertainty quantification used by the global fit.","marker":"[25]"},{"why":"Defines the missing higher-order uncertainty procedure that guards against overinterpreting the chi-square improvement.","marker":"[26]"},{"why":"Provides the comparison global QCD fit with a U(1)_{B-L} Z' boson, which shows no improvement and supports the specificity of the dark photon signal.","marker":"[27]"},{"why":"Shows that O(1) couplings to light dark matter broaden the resonance and relax direct search limits by one to two orders of magnitude.","marker":"[28]"}],"fun_headline_variants":["Dark photon boosts QCD fit by 6.5 sigma","First indirect dark photon hint from DIS data","6.5 sigma preference for dark photon in QCD analysis","Dark photon improves global QCD analysis at 6.5 sigma","New global QCD fit favors dark photon by 6.5 sigma"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central assumption is that the $\\chi^2$ improvement is a genuine new-physics signal rather than an artefact of the parton parametrization and the treatment of theoretical uncertainties in the QCD fit.","fun_headline_variants_meta":{"raw":{"variants":["Dark photon boosts QCD fit by 6.5 sigma","First indirect dark photon hint from DIS data","6.5 sigma preference for dark photon in QCD analysis","Dark photon improves global QCD analysis at 6.5 sigma","New global QCD fit favors dark photon by 6.5 sigma"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000252,"raw_usage":{"total_tokens":1525,"prompt_tokens":871,"completion_tokens":654,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":487,"completion_tokens_details":{"reasoning_tokens":569}},"tokens_in":487,"tokens_out":654,"duration_ms":6727,"temperature":1.0,"reasoning_tokens":569,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:18:40.859672+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A repeat of the global fit with a more flexible parton parametrization, or at next-to-next-to-leading order in the strong coupling, that shows no significant $\\chi^2$ improvement would falsify the claimed preference; a direct search sensitive to a broad resonance in the 4--6 GeV mass range that excludes the quoted $\\epsilon$ window would also do so.","supporting_citations":[{"cited_title":"Constraints on the dark photon from deep inelastic scattering","cited_arxiv_id":"2111.05664","evidence_quote":"Provides the earlier deep-inelastic-scattering constraints on the dark photon that the global analysis extends."},{"cited_title":"Constraints on the $U(1)_{B-L}$ model from global QCD analysis","cited_arxiv_id":"2410.01205","evidence_quote":"Provides the comparison global QCD fit with a U(1)_{B-L} Z' boson, which shows no improvement and supports the specificity of the dark photon signal."}],"review_version":1}