{"id":"665c611d-e4b1-4ce9-aa17-00b6342b5fdc","arxiv_id":"2506.19355","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Benchmarking the GENIE neutrino event generator against semi-exclusive electron-carbon data shows CCQE models overestimate reduced cross sections at missing momenta below about 80 MeV/c.","lead":"This paper compares neutrino-event-generator predictions to electron scattering data on carbon and finds persistent mismatches at low missing momentum. It shows that semi-exclusive electron data can serve as a testing ground for nuclear models used in neutrino oscillation experiments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Low-pm overestimate may be an artifact of mapping GENIE's fixed-binding 1D events to (pm, εm) via Eq. (17); the mapping is unvalidated.","rationale":"The reader's weakest assumption identifies the same load-bearing point: the approximate kinematic mapping from GENIE events to the reduced cross section is not validated, and the paper's own admission that GENIE models are one-dimensional in momentum with fixed binding energies means the shell-resolved εm information in the electron data is lost. My read of the full text confirms that this mapping is essential to every comparison in Figs. 3–7 and to the headline conclusion about pm < 80 MeV/c. The paper provides no control test, no closure check, and no code or configuration files to verify that Eq. (24)'s multiplication by Δεm is equivalent to a proper integration over the experimental missing-energy window. Because the concern is real but not yet demonstrated to be fatal, the appropriate verdict remains CONDITIONAL: the qualitative plots may be correct, but the quantitative overestimate and the 'RDWIA does better' conclusion should not be taken at face value until the mapping is validated. I therefore recommend no change to the reader's verdict.","tokens_in":15526,"tokens_out":14148,"duration_ms":157736,"concrete_test":"Run a control extraction with a generator model that has a two-dimensional spectral function (e.g., GENIE's sf1d or a standalone MC using a Benhar carbon SF). Select the same 1μ1p channel and compute σred(pm) in two ways: (i) with the paper's reconstructed formulas, Eqs. (17)–(22) and the window Δεm; (ii) with truth-level initial-nucleon momentum and removal energy from the event record, histogrammed directly in (pm, εm) and integrated over the experimental missing-energy windows. Repeat the truth-level procedure for the LFG models by using their stored bound-nucleon kinematics. If the reconstructed and truth-level σred agree and the pm < 80 MeV/c excess persists, the concern is refuted; if the excess appears only after the approximate mapping, the central claim is an artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the conversion of GENIE CCQE events into a reduced cross section directly compared with (e,e'p) data. After selecting 1μ1p events, the paper defines the missing energy by Eq. (17), ε̃m = ω − Tp, converts to (pm, εm) via Eqs. (19) and the Jacobian (22), and obtains σred from Eq. (25). The paper itself states in Sec. III that GENIE's nuclear models provide only one-dimensional dσ/dpm with fixed binding energies, so the shell-resolved missing-energy structure in the electron data (10–25 MeV and 30–50 MeV windows) is not preserved. For such a fixed-binding model, the emitted events occupy a narrow εm region near the separation energy, so the distribution is not flat in εm. Yet Eq. (24) as written multiplies a per-MeV density d2σ/dpmdεm by the full experimental window Δεm, implicitly assuming flatness. If the true εm distribution inside the window is not flat, the normalization and the low-pm shape of the extracted σred are distorted. The headline discrepancy is concentrated at pm < 80 MeV/c, and the transformation from measured lepton/proton kinematics into this low-pm region is precisely where the approximate missing-energy definition and Jacobian are least tested. If this mapping biases the low-pm region, the reported GENIE overestimate may be an artifact rather than a property of the nuclear models. A related unvalidated choice is the use of 1p+Nn events rather than 1p+0n events as the analogue of the proton-only electron final states.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes using semi-exclusive electron-carbon scattering data—reduced cross sections and nuclear transparency—as a benchmark for the nuclear models implemented in the GENIE neutrino event generator. The authors simulate charged-current quasielastic neutrino-carbon events with five GENIE model configurations (RFG, LFG, SuSAv2-MEC, spectral function, and effective spectral function), convert the final-state kinematics into a reduced cross section as a function of missing momentum via an approximate missing-energy definition, and compare the results with (e,e'p) data from Tokyo, Saclay, SLAC, and JLab. They report that GENIE models based on the local Fermi gas overestimate the measured reduced cross sections for missing momenta below about 80 MeV/c, while RDWIA calculations agree better. They also compare GENIE nuclear transparency predictions with electron-scattering transparency data and find satisfactory agreement, with some shortcomings attributed to missing multinucleon strength.","tokens_in":15893,"tokens_out":8024,"duration_ms":83966,"significance":"If the comparison is valid, the paper offers a novel, externally grounded test of the nuclear models in a widely used neutrino event generator, using electron-scattering data that were not used to tune those models. The low-missing-momentum region is directly relevant to neutrino energy reconstruction in oscillation experiments, so the claimed discrepancy would be of practical importance. The transparency comparison adds a second, independent validation channel. The paper also demonstrates a general methodology for converting generator events into spectral-function-like observables. However, the central claim depends entirely on the event-to-(pm, epsilon_m) mapping and on matching the missing-energy windows of the experimental data, and the quantitative evidence is reported without a complete goodness-of-fit treatment.","major_comments":[{"comment":"The conversion from GENIE events to the reduced cross section is not sufficiently justified. The paper states in Sec. III that the GENIE nuclear models provide only one-dimensional dsigma/dpm distributions with fixed binding energies, so they cannot produce the shell-resolved d²sigma/dpm/depsilon_m. Nevertheless, Eq. (24) defines dsigma/dpm = Deltaepsilon_m d²sigma/dpm/depsilon_m, which is only correct if d²sigma/dpm/depsilon_m is flat over the window Deltaepsilon_m; for a fixed-binding model the epsilon_m distribution is a narrow peak, not a flat density. Please show explicitly how the event counts are binned and integrated over epsilon_m (for example, by counting events in the full Deltaepsilon_m window and dividing by Delta pm), and state which missing-energy window the electron data actually correspond to. If the data are the shell-separated bins 10-25 MeV and 30-50 MeV mentioned in Sec. III while the GENIE curves are integrated over 80 MeV, the comparison is not apples-to-apples and the reported low-pm overestimate could be an artifact of the window mismatch.","section":"III, Eqs. (17)-(24)"},{"comment":"The quantitative support for the central claim is incomplete. After Fig. 7 the paper reports chi2/DOF approximately 12.6 for the effsf model with SLAC data and approximately 17.5 for the sf1d model with JLab data, but it never gives the number of degrees of freedom, the definition of chi2, the error treatment (statistical errors of the data, normalization uncertainties, or any uncertainty bands on the GENIE predictions), or the pm range used in the comparison. Without this information the phrase \"persistent disagreements\" is not quantitatively established; please provide a defined goodness-of-fit statistic with all components and show the sensitivity of the conclusion to the treatment of data normalization uncertainties.","section":"V, chi-squared paragraph"},{"comment":"The choice of the 1p+Nn event set as the analogue of the proton-only final state in (e,e'p) needs justification. Figure 2 shows that for the G21_11 and spectral-function models the missing-energy distributions of 1p+Nn and 1p+0n events differ noticeably, and the text notes that inelastic FSI changes proton energies and pushes events to higher missing energy. Since the electron data select a knocked-out proton with no associated neutron emission, including events with additional neutrons in the GENIE comparison may bias sigma_red(pm), and the paper does not quantify this bias. Please show the effect of the two selections on the final reduced cross sections, especially at pm below 80 MeV/c, or justify why the 1p+Nn selection is equivalent to the experimental proton-only selection for this observable.","section":"III, event selection (1p+Nn vs 1p+0n)"}],"minor_comments":[{"comment":"The sentence \"The approach presented in this paper provide a great opportunity\" should read \"provides a great opportunity\".","section":"Abstract"},{"comment":"The caption says \"Same as Fig. 3 but for the sf1d (SF) model calculation,\" but the text and figure content describe the effsf model; the caption should be corrected.","section":"Fig. 7 caption"},{"comment":"The Introduction states that results are presented in Sec. IV, but the results actually appear in Sec. V; the cross-reference should be corrected.","section":"Introduction, section numbering"},{"comment":"\"Jacobin\" should be \"Jacobian,\" and the sign convention in the second term of the Jacobian expression should be checked for consistency with the definition of pm in Eq. (20).","section":"Eq. (22)"},{"comment":"\"Femilab\" should be \"Fermilab.\"","section":"Introduction"},{"comment":"\"transparensy\" should be \"transparency.\"","section":"Fig. 9 caption"}],"recommendation":"major_revision","confidential_remarks":"The benchmarking idea is timely and within the journal's scope, and the external electron-scattering data provide a meaningful test of GENIE's nuclear models. The main risk is the kinematic mapping and window matching: if the authors can demonstrate that the event counts are properly integrated over missing energy and that the data windows match the GENIE selection, the central conclusion may survive. I would not recommend rejection at this stage, but the quantitative claims need a rigorous goodness-of-fit treatment and the 1p+Nn versus 1p+0n ambiguity needs to be resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a useful and honest benchmark paper. Butkevich and Luchuk compare five GENIE v3.04 CCQE model configurations (RFG, LFG, SuSAv2, sf1d, effsf) against four semi-exclusive (e,e'p) reduced-cross-section data sets for carbon, using external data that weren't part of GENIE's tuning. The main finding — LFG-based models overestimate the reduced cross section at missing momentum below ~80 MeV/c, while their own RDWIA tracks the data — is visible in the figures and consistent across kinematics. That's a genuinely useful result for the neutrino community, since GENIE is what all the Fermilab experiments use for energy reconstruction, and a persistent low-pm discrepancy in the nuclear model feeds directly into oscillation systematics.\n\nWhat's actually new here is the systematic comparison itself. Earlier generator-electron comparisons used inclusive (e,e') cross sections; this uses the semi-exclusive reduced cross section, which is a stricter test of the nuclear ground state and FSI. The authors also deserve credit for stating plainly that GENIE's nuclear models are one-dimensional in pm with fixed binding energies and cannot reproduce the shell-resolved missing-energy structure in the electron data. That's the right kind of honesty.\n\nThe soft spots are real but mostly don't kill the qualitative conclusion. The kinematic mapping from GENIE events to (pm, εm) — Eq. (17) for missing energy, the Jacobian in Eq. (22), and the window integration — is approximate and, as the stress test notes, unvalidated. The low-pm region is exactly where the 1/pm Jacobian and the experimental acceptance cuts are least benign. I'd push back on one part of the stress test: the flatness concern about Eq. (24) is probably overstated. Read in context, 'integrated over range Δεm' means the density is the window average, so multiplying by Δεm is just the integral — not an assumption of flatness, though the notation is sloppy and should be clarified. The 1p+Nn vs 1p+0n choice is also defensible: electron (e,e'p) experiments don't detect neutrons, so allowing extra neutrons matches the experimental selection.\n\nMore legitimate concerns: χ2/DOF values are quoted without the number of points or the definition of χ2; no configuration files or code are shipped, so reproducing the five model settings is harder than it should be; and the 'RDWIA does better' conclusion rests on the authors' own earlier RDWIA calculations, which are not independently reproduced here.\n\nWho's this for? Anyone working on GENIE validation or neutrino energy reconstruction. It deserves a serious referee — not a desk reject. The referee should push for a validation of the event-to-(pm, εm) mapping (e.g., a closure test against a model with a known spectral function) and for the χ2 details, but the qualitative finding will likely survive.","headline":"A useful and honest GENIE benchmark against semi-exclusive (e,e'p) data with a visible low-pm discrepancy, though the approximate event-to-(pm, εm) mapping and thin quantitative details mean the size of the effect should be treated with caution.","tokens_in":16436,"tokens_out":11031,"would_cite":true,"duration_ms":111056,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["25.30.-c","25.30.Bf","25.30.Pt","13.15.+g"],"model":"deepseek-v4-flash","headline":"This paper shows that electron reduced cross sections can benchmark neutrino generators, and finds GENIE overestimates carbon data below missing momentum 80 MeV/c.","keywords":["GENIE event generator","reduced cross section","nuclear transparency","quasi-elastic neutrino scattering","electron scattering benchmark","distorted spectral function","missing momentum","RDWIA"],"falsifier":"Run the same five GENIE model configurations but reconstruct missing energy with the full recoil term, $\\varepsilon_m = \\omega - T_p - \\varepsilon_B$, instead of $\\tilde{\\varepsilon}_m = \\omega - T_p$, and repeat the comparison; if the low-$p_m$ overestimate disappears or moves to another kinematic region, the central claim fails as stated. A complementary check is to compare generator predictions against shell-tagged $(e,e'p)$ data selected in the $1p_{3/2}$ ($10<\\varepsilon_m<25$ MeV) and $1s_{1/2}$ ($30<\\varepsilon_m<50$ MeV) missing-energy windows, which GENIE's one-dimensional models cannot currently reproduce.","tokens_in":15317,"feed_emoji":"⚛️","tokens_out":10924,"duration_ms":103168,"temperature":0.7,"pith_summary":"Electron and neutrino scattering off nuclei share the same nuclear-physics content: once the lepton vertex and phase space are divided out, the remaining reduced cross section is a distorted spectral function of the target, so precise electron data can test the nuclear models that neutrino event generators rely on. This paper uses that identity to benchmark five GENIE v3 model configurations against four published 12C(e,e'p) reduced-cross-section datasets spanning beam energies from about 0.5 to 2.5 GeV. Across those datasets the paper reports a persistent disagreement: below missing momentum |pm| ≈ 80 MeV/c—the inferred initial momentum of the knocked-out proton—the GENIE models built on a local Fermi gas momentum distribution (G18_10a and G21_11) overestimate the data significantly, while a relativistic distorted-wave impulse approximation (RDWIA) calculation describes the measured cross sections well. If correct, the result means that neutrino energy reconstruction, which depends on these nuclear models, carries an extra systematic uncertainty in exactly the low-missing-momentum region.","feed_headline":"Electron data expose a low-momentum flaw in GENIE's nuclear models","feed_subtitle":"GENIE overestimates carbon reduced cross sections below 80 MeV/c; a distorted-wave model matches the data.","key_machinery":"The load-bearing object is the reduced cross section $\\sigma_{\\mathrm{red}}(p_m) = (d\\sigma/dp_m)/(K\\,\\sigma_{\\ell N})$, the measured or simulated fivefold cross section divided by phase space and by the elementary electron- or neutrino-nucleon cross section; in the nonrelativistic plane-wave limit it becomes the bound-nucleon momentum distribution, and with final-state interactions it is the distorted spectral function. Because this object is nearly independent of the lepton probe, electron data can be used to test the nuclear parts of a neutrino event generator in isolation. The practical machinery is the mapping of generator events onto $(p_m,\\varepsilon_m)$: GENIE's nuclear models only provide one-dimensional $d\\sigma/dp_m$ with fixed binding energies, so the paper selects $1\\mu1p$ events, defines $\\tilde{\\varepsilon}_m = \\omega - T_p$, and uses the Jacobian $J(\\theta_p)$ of Eq. (22) to convert the lepton-proton kinematics into missing momentum and missing energy.","core_discovery":"The paper's central claim is that the reduced cross section $\\sigma_{\\mathrm{red}}(p_m)$ of semi-exclusive $(l,l'p)$ scattering is, up to small Coulomb corrections, the same object for electron and neutrino probes: after dividing out phase space and the elementary lepton-nucleon cross section, what remains is a distorted nuclear spectral function determined by the target and by the ejected nucleon's interaction with the residual nucleus. Using this identity, the paper generates $10^8$ charged-current quasielastic neutrino events per GENIE model configuration, keeps events that end as one muon plus one proton after the intranuclear cascade, maps them to $(p_m,\\varepsilon_m)$ using the approximate missing-energy definition $\\tilde{\\varepsilon}_m = \\omega - T_p$ and the Jacobian of Eq. (22), and compares the resulting $d\\sigma/dp_m$ with four published $^{12}\\mathrm{C}(e,e'p)$ datasets. On that comparison the paper reports persistent disagreements: at missing momenta below about 80 MeV/$c$, the GENIE models based on a local Fermi gas momentum distribution (G18_10a and G21_11) and also the spectral-function-based sf1d and effsf models overestimate the data, while at higher $p_m$ they tend to underestimate; the G18_02a Fermi-gas model shows the opposite pattern at low beam energy. A relativistic distorted-wave impulse approximation (RDWIA) treatment of the same kinematics is found to describe the measured reduced cross sections well, and GENIE's nuclear transparency for carbon is found to be roughly consistent with the transparency data, though the paper concludes that both GENIE and RDWIA need additional strength beyond single-nucleon knockout.","pith_inferences":["If the reported low-momentum overestimate is real, its likely origin is that the LFG and effective spectral-function distributions put too much probability at small initial nucleon momenta; testing the generator's pre-FSI momentum distribution against measured spectral functions would separate that cause from distortions introduced by the cascade.","The paper's comparison discards shell-resolved missing-energy information because GENIE's models are one-dimensional; implementing two-dimensional spectral functions in generators would make the low-$p_m$ comparison much sharper and might resolve whether part of the discrepancy is an artifact of the kinematic mapping.","A practical extension would be to quantify how much the low-$p_m$ overestimate shifts the reconstructed neutrino energy in a simulated oscillation analysis; that number would translate this physics disagreement directly into a systematic uncertainty.","Repeating the benchmark on a heavier target such as argon would show whether the failure is specific to carbon or generic to the momentum-distribution parametrizations used in current generators."],"forward_implications":["Neutrino energy reconstruction in experiments that rely on GENIE will inherit a systematic error in the low-missing-momentum region, where the ejected proton carries little recoil momentum.","The electron-scattering datasets used here provide an out-of-sample test, since they were not used to tune the GENIE configurations; the reported disagreement therefore counts against those nuclear models rather than being a fit artifact.","RDWIA-style unfactorized calculations, which include the distortion of the outgoing nucleon wave function, reproduce the data better and are a natural starting point for improving event-generator nuclear models.","Because GENIE's nuclear transparency predictions agree roughly with data, the failure appears concentrated in the ground-state momentum distribution or quasielastic vertex rather than in the whole intranuclear-cascade description.","The same reduced-cross-section benchmark can be applied to other targets and to other generators, turning precise electron data into a routine validation step for neutrino interaction models."],"supporting_citations":[{"why":"Provides one of the four published carbon (e,e'p) reduced-cross-section datasets used as the benchmark.","marker":"[31, 32]"},{"why":"Provides the low-energy carbon dataset against which the generator predictions are compared.","marker":"[33]"},{"why":"Provides a high-energy carbon dataset that constrains the comparison at large Q^2.","marker":"[34]"},{"why":"Provides a second high-energy carbon dataset, including kinematics close to those of current neutrino experiments.","marker":"[35]"},{"why":"Gives the RDWIA reduced-cross-section calculations that the paper uses as the accurate reference description.","marker":"[23-26]"},{"why":"Identifies the GENIE event generator and its version, the software system under test.","marker":"[27, 28]"},{"why":"Documents the five GENIE model configurations (RFG, LFG, SuSAv2-MEC, SF, effective SF) whose outputs are compared with data.","marker":"[36-38]"},{"why":"Prior test of GENIE's cascade model against nuclear transparency data, used as a comparison point for the transparency part of the study.","marker":"[21]"},{"why":"Supplies the method for selecting 1mu1p events and mapping their kinematics to missing momentum and missing energy.","marker":"[30]"}],"fun_headline_variants":["GENIE fails electron-scattering test at low momenta","Electron data reveal GENIE's low-momentum flaw","Below 80 MeV/c, GENIE overestimates carbon","GENIE vs electrons: low-momentum disagreement exposed","Electron scattering pins down GENIE's weak spot"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The comparison assumes that GENIE events selected as one muon plus one proton and mapped to missing momentum/energy via the approximate formula $\\tilde{\\varepsilon}_m = \\omega - T_p$ and the Jacobian of Eq. (22) reproduce the experimental $(e,e'p)$ reduced cross section; if that kinematic mapping distorts the low-$p_m$ region, the reported overestimate would be an artifact rather than a property of the nuclear models.","fun_headline_variants_meta":{"raw":{"variants":["GENIE fails electron-scattering test at low momenta","Electron data reveal GENIE's low-momentum flaw","Below 80 MeV/c, GENIE overestimates carbon","GENIE vs electrons: low-momentum disagreement exposed","Electron scattering pins down GENIE's weak spot"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1647,"prompt_tokens":1063,"completion_tokens":584,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":679,"completion_tokens_details":{"reasoning_tokens":502}},"tokens_in":679,"tokens_out":584,"duration_ms":6050,"temperature":1.0,"reasoning_tokens":502,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:33:29.143101+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same five GENIE model configurations but reconstruct missing energy with the full recoil term, $\\varepsilon_m = \\omega - T_p - \\varepsilon_B$, instead of $\\tilde{\\varepsilon}_m = \\omega - T_p$, and repeat the comparison; if the low-$p_m$ overestimate disappears or moves to another kinematic region, the central claim fails as stated. A complementary check is to compare generator predictions against shell-tagged $(e,e'p)$ data selected in the $1p_{3/2}$ ($10<\\varepsilon_m<25$ MeV) and $1s_{1/2}$ ($30<\\varepsilon_m<50$ MeV) missing-energy windows, which GENIE's one-dimensional models cannot currently reproduce.","supporting_citations":[{"cited_title":"Mougey, M","cited_arxiv_id":null,"evidence_quote":"Provides the low-energy carbon dataset against which the generator predictions are compared."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides a high-energy carbon dataset that constrains the comparison at large Q^2."},{"cited_title":"Dutta et al","cited_arxiv_id":null,"evidence_quote":"Provides a second high-energy carbon dataset, including kinematics close to those of current neutrino experiments."},{"cited_title":"Dytman, Y","cited_arxiv_id":null,"evidence_quote":"Prior test of GENIE's cascade model against nuclear transparency data, used as a comparison point for the transparency part of the study."},{"cited_title":"Nikolakopoulos, R","cited_arxiv_id":null,"evidence_quote":"Supplies the method for selecting 1mu1p events and mapping their kinematics to missing momentum and missing energy."}],"review_version":2}