{"id":"4bc3785a-f83d-4d4c-aba3-0881e52554e5","arxiv_id":"2501.02852","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The E_G gravity test measured from DES photometric galaxies and Planck CMB lensing is consistent with general relativity, with a forecast that future CSST and CMB-S4 data could reach about 1% precision.","lead":"This paper measures a cosmological statistic called E_G, which tests whether Einstein's general relativity holds on the largest scales, using galaxy positions from the Dark Energy Survey and gravitational lensing maps of the cosmic microwave background from Planck.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fourth-bin consistency claim rests on untested bσ8(z4) bias: E_G(z4)=0.296±0.069 is ~3σ below the ΛCDM prediction unless the DES Y3 bσ8=0.865 is underestimated; the paper flags but does not test this.","rationale":"The reader's CONDITIONAL verdict and its identified weakest assumption (unbiased DES bσ8 inputs) match my independent read; I agree and sharpen the concern to bin 4. The paper applies the Pullen et al. (2015) estimator carefully: jackknife covariances with Hartlap and Percival corrections, conservative scale cuts, shot-noise and window-function corrections, and an honest Section 4.4 discussion of the fourth-bin discrepancy. The problem is that the abstract's central claim (consistent with ΛCDM) is only true for bin 4 under the untested assumption that bσ8(z4) = 0.865 is biased low. Evaluating the paper's own GR prediction E_G = Ωm0/f with its fiducial Ωm0 = 0.336 puts bin 4 about 3σ below the model, so the consistency claim either overstates the data or inherits an unverified systematic. I found no reason to REJECT: bins 1–3 are consistent within about 1.5σ, the paper discloses the bin-4 issue, and the external ACT DR4 result supports the suspected bias direction. I also found no reason to ACCEPT as-is: the quoted values and the abstract's wording need revision pending the check. The ANN uncertainty and magnification-bias points are real but secondary: Section 4.4 shows C_ℓ^gκ contributes more than 90% of the E_G error, so even a several-fold underestimate of σ(fσ8) shifts quoted E_G errors by only a few percent; and the forecast's 1% claim is internally inconsistent with the paper's own 6% magnification systematic, though the conclusions do acknowledge this. The single decisive test is the bσ8(z4) replacement described above, which determines whether the fourth bin is a corrected ΛCDM-consistent point or a genuine tension.","tokens_in":22185,"tokens_out":19146,"duration_ms":161511,"concrete_test":"Recompute E_G(z4) for the 0.7 < z < 0.85 bin using an alternative bσ8(z4) in place of the DES Y3 3×2pt value 0.865 ± 0.034. Two concrete options: (1) use the ACT DR4 MagLim bias from Marques et al. (2024), which the paper cites for the 2.43σ C_ℓ^gg/C_ℓ^gκ discrepancy in this bin; (2) fit the paper's own measured C_ℓ^gg bandpowers in Figure 3 (bin 4) against the Abbott et al. (2023) CAMB/CCL model with σ8 fixed to 0.746 to derive a self-consistent bσ8 — this is an independent check because Figure 3's theory curves use the same DES bias values that enter β. Propagate the new β(z4) = fσ8/bσ8 through Eq. (5) with the same C_ℓ^gκ, Γ, and jackknife covariance, and compare with the ΛCDM prediction E_G = Ωm0/f ≈ 0.51.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — that all four E_G measurements are consistent with ΛCDM — is decided by the fourth bin, and only under an assumption the paper never tests. With the paper's own fiducial cosmology (Ωm0 = 0.336) and the GR expression E_G = Ωm0/f(z) from Section 2.1, the predicted values are approximately 0.55, 0.53, 0.52, and 0.51 at z = 0.30, 0.47, 0.63, and 0.80. The measured values in Eq. (31) fall 1.3σ, 0.9σ, 1.5σ, and 3.1σ below those predictions; using the ANN-derived fσ8 values from Table 2 (f = fσ8/0.746) gives an even larger offset in bin 4, about 4σ. As quoted, therefore, bin 4 is not consistent with ΛCDM at 68% C.L., and the abstract's blanket consistency statement is stronger than the data support. The paper's only resolution is the suggestion in Section 4.4 that bσ8(z4) = 0.865 ± 0.034 from the DES Y3 3×2pt chains is underestimated, which would inflate β and suppress E_G. That explanation is plausible — Marques et al. (2024) report a 2.43σ disagreement between the bias favored by C_ℓ^gg and by C_ℓ^gκ in exactly this bin — but the paper never recomputes E_G with an alternative bias. The assumption is load-bearing in both directions: if bσ8(z4) is unbiased, the fourth bin is a genuine ~3σ tension with ΛCDM and the central claim fails as stated; if bσ8(z4) is biased as suspected, the quoted E_G(z4) = 0.296 is itself the wrong number. Either way the fourth-bin result, and therefore the headline consistency claim, is not robust until this is resolved. Secondary issues — ANN fσ8 errors of ~2.5% derived from input measurements with 10–20% errors, and a forecast 1% precision that sits atop the paper's own ~6% magnification-bias floor — would alter error budgets but not the central measurement; the untested bσ8(z4) assumption is the load-bearing one.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper measures the E_G statistic, a ratio of gravitational lensing to galaxy clustering, as a test of general relativity on cosmological scales. The authors use the DES Y3 MagLim photometric galaxy sample and the Planck 2018 CMB lensing map, estimating the angular power spectra C_ℓ^gg and C_ℓ^gκ with the NaMaster pseudo-Cℓ estimator. To handle the photometric-redshift limitation, they introduce a new RSD parameter β = fσ8/bσ8, where fσ8 is reconstructed from 66 spectroscopic growth-rate measurements using an artificial neural network (ReFANN) and bσ8 is taken from the DES Y3 3×2pt chains. They obtain E_G = 0.354 ± 0.146, 0.452 ± 0.092, 0.414 ± 0.069, and 0.296 ± 0.069 in four redshift bins centered at z = 0.30, 0.47, 0.63, and 0.80, and claim consistency with ΛCDM. They also forecast E_G constraints for the CSST photometric survey combined with CMB-S4 lensing, projecting ~1% precision.","tokens_in":22637,"tokens_out":2550,"duration_ms":26433,"significance":"If the measurements are robust, the paper provides new E_G constraints from a photometric survey combined with CMB lensing, extending the redshift range of E_G tests and demonstrating a novel ANN-based treatment of the RSD parameter. The use of public codes, conservative scale cuts, and Hartlap/Percival covariance corrections are strengths, as is the explicit acknowledgment of the fourth-bin bias issue in Section 4.4. However, the headline consistency claim is sensitive to the fourth redshift bin, where the measured E_G is about 3σ below the ΛCDM prediction under the paper's own fiducial cosmology. The paper's suggested explanation (underestimated bσ8 at z≈0.80) is plausible but untested, and the ANN-derived fσ8 errors appear implausibly small given the input data errors. These issues, if resolved, would make the paper a useful contribution; as it stands, the central claim is not yet fully supported.","major_comments":[{"comment":"The fourth-bin measurement E_G(z4) = 0.296 ± 0.069 is approximately 3σ below the ΛCDM prediction computed from the paper's own fiducial cosmology (Ω_m0 = 0.336, f(z) from GR), as the text itself notes in §4.4. The only resolution offered is the suggestion that bσ8(z4) = 0.865 ± 0.034 from the DES Y3 chains is underestimated, which would inflate β and suppress E_G. This explanation is not tested: the paper never recomputes E_G with an alternative bias value or adds a systematic error to absorb the discrepancy. The abstract's blanket statement that all four measurements 'are consistent with the predictions of the standard ΛCDM model' is therefore not supported by the presented analysis. The authors should either provide a quantitative test of the bias hypothesis or qualify the consistency claim to exclude or downgrade bin 4.","section":"§4.4 and Eq. (31)"},{"comment":"The ANN-reconstructed fσ8 values have quoted uncertainties of ~0.010–0.015, which are much smaller than the typical errors of the 66 input measurements (many are 0.03–0.18, see Table 1). The paper gives no description of how these uncertainties are computed (e.g., bootstrap, network variance, or covariance with the input data) and no validation that the ANN error estimate is unbiased. Since β = fσ8/bσ8 and the E_G estimate depends linearly on β, an underestimated fσ8 error would propagate into underestimated β and E_G errors, potentially affecting the significance of the consistency claims. Please provide details of the ReFANN error estimation and a comparison with a more standard reconstruction (e.g., a Gaussian process with explicit prior sensitivity checks).","section":"Table 2 and §4.1.1"},{"comment":"The calibration factor Γ(ℓ,z) is computed using theoretical C^{mg}_ℓ and Q^{mg}_ℓ that depend on the fiducial ΛCDM parameters (Ω_m0, σ8, h, etc.) and on the assumed linear matter power spectrum. The paper states that E_G is independent of galaxy bias and σ8, but this only holds if the calibration factor is sufficiently precise and if the bias and growth inputs are mutually consistent. The bσ8 values are taken from DES Y3 3×2pt chains that assume a particular background cosmology and lensing kernel, so the claimed 'model-independent' status of the bσ8 input is overstated. The paper should state whether the calibration factor's theoretical uncertainty is propagated into the E_G error budget, and if not, justify why it is negligible.","section":"§2.3, Eqs. (11)–(15)"}],"minor_comments":[{"comment":"The phrase 'which are consistent with the predictions of the standard ΛCDM model' should be made conditional on the fourth-bin caveat, or the conclusion should be rephrased to 'consistent within the current large uncertainties, with the possible exception of the highest-redshift bin.'","section":"Abstract and §6"},{"comment":"The effective redshifts are quoted to two decimal places, but the bin edges are given as 0.20–0.40, 0.40–0.55, etc. It would be clearer to state whether the effective redshift is computed from the weighted n(z) and to provide the redshift distribution of each bin in a table or figure with the effective redshifts marked.","section":"§4.1.1, Table 2"},{"comment":"The right panels show R_ℓ = C^{gκ}_ℓ / C^{gg}_ℓ, but the ordinate ranges for the four bins differ by orders of magnitude. Please use the same scale or explicitly note the different y-axis limits to avoid visual overinterpretation of the fluctuations.","section":"Figure 3"},{"comment":"The forecast covariance formula uses the standard Gaussian expression, but the shot-noise and beam terms for the CMB-S4 lensing map should be specified more precisely (particularly the N^{κκ}_ℓ term) so that the forecast is reproducible.","section":"§5.3, Eq. (38)"},{"comment":"There are several typographical errors, including 'preform' for 'perform' in §4.1.1 and inconsistent use of 'EG' vs 'E_G' in the text and figures. A careful proofreading pass is needed.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and presents a plausible method for E_G estimation with photometric surveys. However, the central consistency claim rests on the fourth redshift bin, and the paper's own admission of a possible bias in that bin, without a test, makes the claim fragile. The ANN fσ8 error estimates are also a red flag that should be examined before publication. I recommend major revision with a request to either resolve the fourth-bin issue or substantially weaken the abstract and conclusions. The forecast section is more speculative and could be shortened, but it is not the main concern."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper before you delegate it: it is a legitimate application of the existing E_G estimator to DES MagLim plus Planck lensing, producing four new E_G measurements that are consistent with GR within the uncertainties — except possibly the fourth bin, where the paper's own consistency claim is on thin ice.\n\nWhat is actually new: the numbers in Eq. (31), i.e., E_G = 0.354 ± 0.146, 0.452 ± 0.092, 0.414 ± 0.069, 0.296 ± 0.069 at z = 0.30, 0.47, 0.63, 0.80 for the DES MagLim sample combined with Planck 2018 lensing. The methodology is taken from Pullen et al. (2015) and Wenzl et al. (2024), including the β = fσ8/bσ8 formulation — which is an algebraic identity with f/b, not a new idea, and the paper's wording \"new definition\" is overstatement. The mass of the paper is a careful application: NaMaster pseudo-Cℓ, jackknife covariances with Hartlap and Percival corrections, conservative scale cuts, and a reasonable treatment of magnification bias as a systematic. That part is solid.\n\nNow the soft spots. The main issue is the fourth bin. The stress-test is right: with the paper's own fiducial cosmology, the ΛCDM prediction at z = 0.80 is about 0.51, and the measured 0.296 ± 0.069 is ~3σ low. The paper notices and suggests bσ8(z4) from the DES chains is underestimated, citing Marques et al. (2024). But it never tests that hypothesis by recomputing E_G with a different bias. If bσ8(z4) is unbiased, the fourth bin is a real tension and the abstract's blanket consistency statement fails. If bσ8(z4) is biased as suspected, then the quoted E_G(z4) is itself the wrong number. Either way, the headline result is not robust until this is resolved. The authors should be required to either drop the fourth bin, rerun with an alternative bias, or explicitly present the result as conditional. Second, the ANN-derived fσ8 errors in Table 2 are implausibly small: ~0.01 on fσ8 from input measurements with ~0.05–0.1 errors. The ANN output error should account for scatter in the training data; as reported, it looks like an artifact of the regression. This doesn't change the E_G values much, but it understates the error budget. Third, the forecast of ~1% precision sits on top of the paper's own 6% magnification-bias estimate, so the claim \"achievable constraint of approximately 1%\" needs a caveat about unmodeled systematics.\n\nOverall: this is a useful data point, not a breakthrough. The measurement methodology is sound; the presentation is honest about the fourth-bin issue, but the paper's own claim goes beyond what it demonstrates. A serious referee should require the bσ8(z4) resolution and a softening of the abstract before publication. I would send it to review with a request for major revision.","headline":"New E_G measurements from DES MagLim plus Planck lensing are broadly consistent with GR, but the fourth bin sits ~3σ low and the paper's consistency claim rests on an untested bias assumption.","tokens_in":23277,"tokens_out":2769,"would_cite":true,"duration_ms":24059,"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":"General relativity passes a new DES-Planck lensing test","keywords":["E_G statistic","general relativity","modified gravity","weak gravitational lensing","CMB lensing","galaxy clustering","photometric redshift surveys","large-scale structure"],"falsifier":"Measure $b\\sigma_8$ at $z \\approx 0.8$ from an independent probe, for example the MagLim auto-correlation alone or ACT DR4 lensing with cosmology held fixed; if it comes out significantly above $0.865 \\pm 0.034$, the fourth-bin $E_G$ rises and may land above the $\\Lambda$CDM prediction, directly testing the paper's bias-underestimation explanation.","tokens_in":21935,"feed_emoji":"🌌","tokens_out":8258,"duration_ms":66736,"temperature":0.7,"pith_summary":"The paper sets out to test whether general relativity holds on cosmological scales by measuring the $E_G$ statistic, which compares the gravitational-lensing signal of matter with the growth of cosmic structure, using photometric galaxies from the Dark Energy Survey and the Planck 2018 CMB lensing map. Because photometric redshifts are imprecise, the authors reconstruct the growth rate $f\\sigma_8$ with an artificial neural network and take the galaxy bias $b\\sigma_8$ from DES chains, forming the redshift-space-distortion parameter $\\beta = f\\sigma_8/b\\sigma_8$. They report $E_G = 0.354 \\pm 0.146$, $0.452 \\pm 0.092$, $0.414 \\pm 0.069$, and $0.296 \\pm 0.069$ at $z = 0.30$, $0.47$, $0.63$, and $0.80$, all consistent with $\\Lambda$CDM. A forecast with the China Space Station Telescope and CMB-S4 indicates $E_G$ could be measured to roughly 1% precision, which would separate general relativity from several modified-gravity models.","feed_headline":"General relativity passes a new DES-Planck lensing test","feed_subtitle":"Four redshift bins stay consistent with ΛCDM; future CSST and CMB-S4 data could reach 1% precision.","key_machinery":"The central object is the $E_G$ statistic introduced by Zhang et al. (2007), estimated here as $E_G(\\ell,\\bar{z}) = \\Gamma(\\bar{z})\\, C_\\ell^{g\\kappa} / [\\beta(\\bar{z})\\, C_\\ell^{gg}]$, where $C_\\ell^{g\\kappa}$ is the galaxy-CMB lensing cross-power spectrum, $C_\\ell^{gg}$ the galaxy auto-power spectrum, and $\\beta$ the redshift-space-distortion parameter. The paper's operational move is $\\beta = f\\sigma_8/b\\sigma_8$: the numerator comes from an ANN fit (ReFANN) to 66 spectroscopic growth-rate measurements, and the denominator from model-independent DES Y3 chains. The calibration factor $\\Gamma(z)$ corrects for broad redshift distributions, the lensing kernel, and scale-dependent bias, while jackknife resampling supplies the covariance with Hartlap and Percival corrections for the inverse covariance.","core_discovery":"The central claim is that a photometric-redshift survey can deliver meaningful $E_G$ constraints once the redshift-space-distortion parameter is rewritten as $\\beta = f\\sigma_8/b\\sigma_8$, with $f\\sigma_8$ reconstructed from the existing spectroscopic growth-rate compilation and $b\\sigma_8$ taken from DES Y3 chains. Applying this to four MagLim tomographic bins yields $E_G = 0.354 \\pm 0.146$, $0.452 \\pm 0.092$, $0.414 \\pm 0.069$, and $0.296 \\pm 0.069$ at $z = 0.30$, $0.47$, $0.63$, $0.80$, consistent with $\\Lambda$CDM and with earlier estimates. The paper also claims that the same pipeline applied to future CSST and CMB-S4 data will reach roughly 1% precision, enough to distinguish general relativity from chameleon and $f(R)$ models.","pith_inferences":["If the fourth-bin bias is indeed underestimated, an independent $b\\sigma_8$ measurement from the MagLim auto-correlation alone or from ACT lensing would shift $E_G(z\\approx0.8)$ upward, providing a direct test of the paper's explanation.","The ANN reconstruction inherits the selection of the 66-point growth-rate compilation; swapping in a stricter or updated RSD sample would quantify how much of the result depends on that choice.","The same $\\beta$-split pipeline transfers naturally to other photometric surveys such as LSST and Euclid, where magnification-bias corrections will matter at the few-percent level.","Combining future CMB-S4 lensing with CSST clustering may also break the current degeneracy between galaxy bias and growth in the highest tomographic bin."],"forward_implications":["General relativity remains consistent with the combined DES and Planck lensing data at redshifts 0.30 to 0.80, with no significant scale dependence in $E_G$.","Photometric surveys can test gravity using the $\\beta = f\\sigma_8/b\\sigma_8$ split instead of direct redshift-space-distortion measurements.","The low $E_G$ in the highest redshift bin likely reflects an underestimated $b\\sigma_8$ from the DES chains rather than new physics; a corrected bias would move it toward the $\\Lambda$CDM prediction.","Future CSST and CMB-S4 data could reach roughly 1% precision on $E_G$, separating general relativity from chameleon gravity at the 5$\\sigma$ level and from $f(R)$ gravity at the 13$\\sigma$ level for $B_0 > 10^{-7}$.","Magnification bias shifts the forecast $E_G$ by up to roughly 6%, so future high-precision analyses must include it explicitly."],"supporting_citations":[{"why":"Defines the $E_G$ statistic as the ratio of the gravitational-potential Laplacian to the velocity divergence, giving the paper its central probe.","marker":"Zhang et al. (2007)"},{"why":"Proposes the CMB-lensing version of $E_G$ and the estimator $E_G = \\Gamma C^{g\\kappa}/(\\beta C^{gg})$ used throughout.","marker":"Pullen et al. (2015)"},{"why":"Supplies the DES Y3 3x2pt chains from which the $b\\sigma_8$ values and the adopted cosmology are taken.","marker":"Abbott et al. (2023)"},{"why":"Provides the Planck 2018 minimum-variance CMB lensing convergence map used for $C_\\ell^{g\\kappa}$.","marker":"Aghanim et al. (2020)"},{"why":"Introduces the $\\beta = f\\sigma_8/b\\sigma_8$ decomposition and the analytic shot-noise correction adopted here.","marker":"Wenzl et al. (2024)"},{"why":"One source of the 66 compiled $f\\sigma_8$ measurements that feed the ANN reconstruction.","marker":"Kazantzidis & Perivolaropoulos 2018"},{"why":"Provides the ReFANN package used to reconstruct $f\\sigma_8(z)$ from the growth-rate compilation.","marker":"Wang et al. (2020)"},{"why":"Supplies the calibration-factor treatment, nonlinear scale cuts, and the $\\beta$-error relation used in the forecast.","marker":"Yang & Pullen (2018)"},{"why":"Gives the magnification-bias parameters and the comparison showing the auto-correlation favors a higher bias than galaxy-CMB lensing in the $z\\approx0.8$ bin.","marker":"Marques et al. (2024)"}],"fun_headline_variants":["EG from DES and Planck: GR consistent with ΛCDM","Photometric survey tests gravity with CMB lensing","Future CSST+CMB-S4 could test GR at 1% level","New EG measurement: Einstein holds at cosmological scales"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that the $b\\sigma_8$ values taken from the DES Y3 3x2pt chains are unbiased in every tomographic bin, especially the highest one, where the chains give an unusually low value.","fun_headline_variants_meta":{"raw":{"variants":["EG from DES and Planck: GR consistent with ΛCDM","Photometric survey tests gravity with CMB lensing","Future CSST+CMB-S4 could test GR at 1% level","New EG measurement: Einstein holds at cosmological scales"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000614,"raw_usage":{"total_tokens":2960,"prompt_tokens":1156,"completion_tokens":1804,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":772,"completion_tokens_details":{"reasoning_tokens":1734}},"tokens_in":772,"tokens_out":1804,"duration_ms":18059,"temperature":1.0,"reasoning_tokens":1734,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:02:32.307654+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure $b\\sigma_8$ at $z \\approx 0.8$ from an independent probe, for example the MagLim auto-correlation alone or ACT DR4 lensing with cosmology held fixed; if it comes out significantly above $0.865 \\pm 0.034$, the fourth-bin $E_G$ rises and may land above the $\\Lambda$CDM prediction, directly testing the paper's bias-underestimation explanation.","supporting_citations":[{"cited_title":"2007, Physical Review Letters, 99, 141302","cited_arxiv_id":null,"evidence_quote":"Defines the $E_G$ statistic as the ratio of the gravitational-potential Laplacian to the velocity divergence, giving the paper its central probe."},{"cited_title":"R., Alam, S., & Ho, S","cited_arxiv_id":null,"evidence_quote":"Proposes the CMB-lensing version of $E_G$ and the estimator $E_G = \\Gamma C^{g\\kappa}/(\\beta C^{gg})$ used throughout."},{"cited_title":"2023, Physical Review D, 107, 083504","cited_arxiv_id":null,"evidence_quote":"Supplies the DES Y3 3x2pt chains from which the $b\\sigma_8$ values and the adopted cosmology are taken."},{"cited_title":"2020, Astronomy & Astrophysics, 641, A8","cited_arxiv_id":null,"evidence_quote":"Provides the Planck 2018 minimum-variance CMB lensing convergence map used for $C_\\ell^{g\\kappa}$."},{"cited_title":"2024, Physical Review D, 109, 083540","cited_arxiv_id":null,"evidence_quote":"Introduces the $\\beta = f\\sigma_8/b\\sigma_8$ decomposition and the analytic shot-noise correction adopted here."},{"cited_title":"2018, Physical Review D, 97, 103503","cited_arxiv_id":null,"evidence_quote":"One source of the 66 compiled $f\\sigma_8$ measurements that feed the ANN reconstruction."},{"cited_title":"2020, The Astrophysical Journal Supplement Series, 246, 13","cited_arxiv_id":null,"evidence_quote":"Provides the ReFANN package used to reconstruct $f\\sigma_8(z)$ from the growth-rate compilation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the calibration-factor treatment, nonlinear scale cuts, and the $\\beta$-error relation used in the forecast."},{"cited_title":"2024, Journal of Cosmology and Astroparticle Physics, 2024, 033","cited_arxiv_id":null,"evidence_quote":"Gives the magnification-bias parameters and the comparison showing the auto-correlation favors a higher bias than galaxy-CMB lensing in the $z\\approx0.8$ bin."}],"review_version":1}