66 of 66 reviewed claims · counts describe this ledger, not model accuracy.
Original: supportedID-01 · gpt-5.6-sol · grok-4.6 · fable-5.1
The two-cloud spin-2 mechanism is established prior art
Different cloud SEDs and non-collinear Stokes vectors can generate a frequency-dependent resultant angle. Common SEDs factor out. The mechanism and optical-tomography motivation predate this corpus.
Scientific value: Sound organizing principle; literature synthesis rather than new physics.
Original passages (9)
Original: qualifiedID-02 · gpt-5.6-sol · fable-5.1
A scalar polarization SED does not alone fix polarization fraction
Pν=fνP0 with positive f fixes angle. It fixes p=|P|/I only if total intensity has the same scaling. That intensity condition must be stated as part of the model, rather than inferred from Q/U alone.
Scientific value: Separates two measurable observables and prevents a false diagnostic.
Original passages (2)
Original: incorrectID-03 · grok-4.6
Ordinary ILC does not impose a single-dust null
Ordinary ILC minimizes variance subject to unit CMB response. Dust-SED nulling is an additional constrained-ILC condition; non-MBB dust does not automatically bias every ILC implementation.
Scientific value: Material distinction for selecting and implementing component-separation baselines.
Original passages (1)
Original: incorrectID-04 · grok-4.6
The independent Grok layer amplitude has inconsistent units
A_i^P is defined as p0 cos²γ τ, then multiplied by a dimensionless reference-normalized SED to obtain Q/U. Include Bν0(T) in the reference amplitude or use an unnormalized Planck function.
Scientific value: Prevents a physically incorrect forward-model implementation.
Original passages (1)
Original: incorrectID-05 · grok-4.6
A phenomenological MBB is falsifiable
The claim that a phenomenological MBB is not falsifiable is false: additional spectral measurements can reject its predictions. Greater physical detail is not a prerequisite for falsifiability.
Scientific value: Clarifies how to evaluate simple and complex scientific models fairly.
Original passages (1)
Original: incorrectID-06 · fable-5.1
The GPT-driven Fable temperature derivative is reversed
At answer.md:64 the ratio slope has the wrong temperature sign. For ρ=P2 f2/(P1 f1), d lnρ/d lnν=Δβ+g(x1)−g(x2), since d lnBν/d lnν=3−g(x). Equal-temperature examples cannot catch this error.
Scientific value: Affects the direction of predicted frequency rotation and needs an unequal-temperature verification case.
Original passages (1)
Original: qualifiedID-07 · fable-5.1
Perpendicular cloud polarizations can undergo a 90-degree flip
The statement of no rotation at a 90° cloud separation applies while the amplitude ratio remains on one side of unity. Crossing unity passes through zero polarization, where the angle is undefined, then flips the angle by 90°.
Scientific value: Identifies a singular limit for simulations and inference.
Original passages (1)
Original: incorrectID-08 · fable-5.1
Grain growth is not shown to lower beta by the cited model
The essay attributes a lower β to accretion/coagulation in Köhler et al. Their model instead raises β from 1.5 to 1.8 with accretion and to about 2 with coagulation. This is not a universal law for all grains.
Scientific value: A proposed grain-evolution prior has the opposite sign to its cited support.
Original passages (1)
Original: supportedID-09 · gpt-5.6-sol
Grok-driven GPT correctly repairs the Rayleigh-limit argument
Submicron interstellar grains remain in the Rayleigh limit at CMB wavelengths; growth does not generically flatten β through geometric-optics greying. Material and aggregate structure matter.
Scientific value: An evidence-based improvement over the driver brief.
Original passages (1)
Original: incorrectID-10 · fable-5.1
The Zelko paper has four authors
The final explicit correction to a two-author paper is itself wrong. The arXiv record lists Ivan Zelko, Douglas Finkbeiner, Albert Lee and Gregory Green.
Scientific value: Repairs source provenance without disputing the existence of 3D temperature modeling.
Original passages (1)
Original: incorrectID-11 · gpt-5.6-sol
The linked MNRAS paper is by McBride, Bull and Hensley
The final reference labels MNRAS 519,4370 as Sponseller & Kogut. The linked article is McBride, Bull & Hensley (2023); its topic is relevant.
Scientific value: Bibliographic attribution error, distinct from a false scientific result.
Original passages (1)
Original: incorrectID-12 · grok-4.6
Ritacco citation journal locator is wrong
The bibliography lists A&A 672,A22. The primary record gives A&A 670,A163 for the linked Ritacco et al. work.
Scientific value: Improves reliable source navigation.
Original passages (1)
Original: incorrectID-13 · fable-5.1
BK18 decorrelation marginalization is an extension
The baseline fixes no decorrelation. Appendix E.2 tests a variable decorrelation parameter; it is not marginalized in the main baseline as stated.
Scientific value: Avoids misrepresenting the actual published tensor likelihood.
Original passages (1)
Original: supportedID-14 · gpt-5.6-sol · fable-5.1 · grok-4.6
Broad-area nulls and selected-sightline evidence are compatible
Planck broad-area correlation limits, Pelgrims selected-sightline angle differences and SPIDER regional SED measurements test different statistics and selections; they are not mutually contradictory.
Scientific value: A useful reconciliation, with no universal decorrelation amplitude implied.
Original passages (5)
Original: supportedID-15 · fable-5.1 · gpt-5.6-sol · grok-4.6
The quoted recent references are real
Guillet 2510.18305, McCallum 2607.07451 and Tram 2606.13123 are real primary records. Their results concern residual covariance, model-based H I/extinction reconstruction, and optical grain alignment respectively, not a validated CMB-cleaning pipeline.
Scientific value: Extends the literature inventory while respecting each study’s domain.
Original passages (3)
Original: qualifiedID-16 · fable-5.1 · gpt-5.6-sol · grok-4.6
Local angle correlation is not automatically harmonic BB decorrelation
Local cos(2δψ), a Gaussian-index ensemble coefficient and a harmonic Rℓ require different assumptions. Angular covariance, amplitude correlations and E/B mixing must be specified before transferring results to a BB spectrum.
Scientific value: A central safeguard for meaningful simulations and numerical interpretation.
Original passages (5)
Original: qualifiedID-17 · fable-5.1 · gpt-5.6-sol · grok-4.6
One minus correlation squared is an optimized scalar residual
Cres(a)=Cxx(1−R²)+Cyy(a−Cxy/Cyy)². Thus 1−R² applies at the optimal scalar regression coefficient. Knowing R does not by itself remove this residual or specify an r bias.
Scientific value: Provides an exact reusable benchmark and exposes calibration penalties.
Original passages (5)
Original: supportedID-18 · fable-5.1 · gpt-5.6-sol · grok-4.6
The worked two-cloud examples largely reproduce
Independent Planck-function calculations reproduce the equal-temperature examples, the 0.04495 thermodynamic 150/353 amplitude scale and the 0.7% uncorrelated-layer residual under the stated covariance model. These are illustrative calculations rather than measured sky statistics.
Scientific value: Useful educational and numerical reference cases; not independent novelty.
Original passages (3)
Original: unsupportedID-19 · fable-5.1 · grok-4.6
Dust power fractions do not determine an experimental r bias
Quoted residual-r levels are not established by local vector examples alone. A forecast needs spatial fields, E/B transformation, tensor template, mask, bandpasses, noise and the foreground/r likelihood. The order-of-magnitude concern is plausible.
Scientific value: Separates a defensible risk illustration from a validated experiment forecast.
Original passages (5)
Original: qualifiedID-20 · fable-5.1
Cloud angles help conditioning only in the appropriate parameterization
With independent Q and U amplitudes the mixing matrix is I2⊗F and true cloud angles do not change its singular values. With known angles and one real amplitude per cloud, angle-dependent columns can help. Band counts alone do not prove identifiability.
Scientific value: A concrete research extension: compare uncertain angle priors using singular values and posterior coverage.
Original passages (2)
Original: qualifiedID-21 · fable-5.1 · gpt-5.6-sol
Optical-to-emission transfer is an uncertain prior
Starlight and dust-emission polarization orientations differ by 90°. An empirical R_P/p transfer must respect that convention and matched dust columns; alignment and inclination do not cancel exactly under arbitrary population or environment differences.
Scientific value: Preserves the value of tomography while identifying transfer-model uncertainty.
Original passages (3)
Original: incorrectID-22 · fable-5.1
A universal three-layer and 500-pc cutoff is unjustified
The statement that all high-latitude dust lies within about 500 pc in at most three layers is not a valid completeness assumption. Tomography detects more distant structures, and complexity studies find effects from dust beyond the nearest few hundred parsecs.
Scientific value: Prevents missing-layer model bias and motivates an unresolved/far-dust component.
Original passages (1)
Original: incorrectID-23 · fable-5.1
Monochromatic dust weights are not bolometric luminosity weights
τ T^(4+β) is an integrated modified-blackbody scaling, not the fixed-band emission weight. At a chosen frequency use τ(ν/ν0)^β Bν(T).
Scientific value: Prevents overly steep temperature weighting of dust layers.
Original passages (1)
Original: qualifiedID-24 · fable-5.1 · grok-4.6
Temperature and beta dominance must be inferred jointly
Derivative sizes alone do not rank physical contributions without parameter variances and covariances. A fixed-β 15→25 K range changes 353/217 by 11.38%, not 15%. SRoll2 temperature-only compatibility is not proof of universal temperature dominance.
Scientific value: Supports testing both hypotheses rather than imposing a grain-property narrative.
Original passages (2)
Original: incorrectID-25 · grok-4.6
The requested power precision is not derived consistently
The stated 0.009 µK² dust and 7×10⁻5 µK² tensor powers imply about 0.78%, not 0.01%, for their ratio. A desired fractional tensor bias must be specified to turn that into an allowable dust residual.
Scientific value: Makes experimental requirements traceable to a stated error budget.
Original passages (1)
Original: incorrectID-26 · grok-4.6
Misalignment does not put net p below every layer’s p
For unequal layer fractions, p_total is bounded by the intensity-weighted no-cancellation fraction and the largest p_i; it need not lie below the smallest p_i. Common SEDs can still factorize despite misalignment.
Scientific value: Clarifies the observable signature of geometric depolarization.
Original passages (2)
Original: qualifiedID-27 · fable-5.1
Simulated moment-method gains are configuration-specific
Vacher’s σr=8.8×10⁻4 is a specified LiteBIRD BB cross-spectrum simulation result. Higher-frequency improvements and large SO model-mismatch biases in other cited studies likewise depend on their simulated sky and inference setup.
Scientific value: Useful baselines, not universal performance floors or measured biases.
Original passages (3)
Original: qualifiedID-28 · gpt-5.6-sol · fable-5.1
Reference-normalized temperature moments need a reference term
For amplitudes defined at ν0, the first temperature basis is ∂T ln[Bν/Bν0]. Writing only ∂T lnBν can be consistent with a redefined zeroth moment but requires a declared amplitude convention.
Scientific value: Improves equation-to-code traceability without discarding the valid moment approach.
Original passages (2)
Original: proposalID-29 · gpt-5.6-sol · grok-4.6 · fable-5.1
A geometry-informed foreground model remains a research proposal
Test whether uncertain extinction, grain and angle information improves held-out microwave predictions and calibrated r recovery over layered, moment and ILC baselines. The essays do not supply the necessary data analysis or coverage study.
Scientific value: Potential methodological contribution if independently benchmarked; novelty is not demonstrated by combining the ingredients.
Original passages (9)
Original: incorrectdialogue-pelgrims-selection · fable-5.1
Pelgrims target selection explicitly used ≥60° misalignment
G1 line318 removes the threshold because it is absent from the abstract. The full paper §4 and Table1 include ≥60° in both target implementations. Absence from an abstract is not evidence against a method detail.
Scientific value: Restores reproducible target/control selection; not a new discovery.
Original passages (1)
Original: unsupporteddialogue-pelgrims-sparsity · grok-4.6 · fable-5.1
Broad null plus selected detection does not measure a sparse unsafe sky fraction
A broad power-spectrum null and selected angle excess use different statistics, masks and sensitivities. They do not determine what percentage of sky is safe at a target r.
Scientific value: Motivates a selection-aware risk-map calibration instead of a fixed sky fraction.
Original passages (2)
Original: incorrectdialogue-los-exclusive · grok-4.6 · fable-5.1
Chromatic angles and E/B statistics are not exclusive to LOS mixing
Beam averaging of different sightlines also mixes SEDs and polarization directions. Spatially varying multiplication does not commute with harmonic E/B decomposition. These observables alone do not locate mixing along distance.
Scientific value: Separate distance, beam, and analysis-scale averaging in tests.
Original passages (2)
Original: qualifieddialogue-angle-rank · grok-4.6 · fable-5.1
Known two-layer angles help only away from both 0° and 90° degeneracies
For two real amplitudes detX=sin(2Δψ); the best equal-noise geometry is45°, while0° and90° are singular. This dust-only algebra excludes an additional unconstrained single-band CMB Q/U. Merely distinct angles do not guarantee good conditioning.
Scientific value: Provides an exact control case for identifiability simulations.
Original passages (4)
Original: incorrectdialogue-one-band-impossibility · grok-4.6
Two known noncollinear dust angles can separate two reference-band amplitudes
The critic says one frequency can never separate the layers. With exactly known noncollinear directions, two dust-only Stokes equations identify two real amplitudes. Multi-band CMB separation and angle uncertainty are additional problems.
Scientific value: Prevents throwing away legitimate ancillary-angle information.
Original passages (1)
Original: incorrectdialogue-split-error · fable-5.1
Matched reference-band splitting suppresses split errors, not all SED errors
G1 lines223–239 correctly writes an identity but calls every term a product of small quantities. ΣP_i δg_i is first order in SED error; P_i is not an error. The split-only term is contrast-suppressed, conditional on an exact reference sum.
Scientific value: A useful conditional error-budget identity; no universal lenient tolerance or r guarantee.
Original passages (1)
Original: qualifieddialogue-cmilc · gpt-5.6-sol · fable-5.1 · grok-4.6
cMILC spectral nulling differs from fitting spatial moment templates
cMILC nulls frequency vectors such as SED derivatives. A spatial-template phase error affects a template regression, but does not invalidate an exact spectral null. A physically informed spatial regression is a separate estimator requiring validation.
Scientific value: Clarifies the algorithm to implement and avoids claiming pre-existing cMILC as a new method.
Original passages (6)
Original: incorrectdialogue-ordinary-ilc · grok-4.6
Ordinary ILC does not null a specified dust SED
G5 line135 says standard ILC nulls a dust SED. Ordinary ILC constrains unit CMB response and minimizes variance. Nulling dust is an additional constrained-ILC condition. G6 synthesis repairs this point.
Scientific value: Affects baseline comparison and residual interpretation.
Original passages (1)
Original: supporteddialogue-faraday · gpt-5.6-sol
Faraday rotation does not directly measure dust-layer inclination
RM is an electron-density-weighted integral of the LOS magnetic field. Without transverse magnitude, phase association and distance information, it cannot independently yield a dust layer inclination.
Scientific value: Prevents a geometry proxy from being treated as a fully observed amplitude factor.
Original passages (2)
Original: incorrectdialogue-optical-sign · fable-5.1
Optical-to-emission Stokes transfer requires the 90° sign reversal
G2 line151 directly multiplies optical q,u by a positive emission/extinction ratio. Under a common Stokes convention the emission EVPA is perpendicular to optical extinction polarization, giving a minus sign in Q+iU. The original field-angle formula elsewhere does not fix this direct-transfer statement.
Scientific value: Essential for a correct layer template and signed angle prediction.
Original passages (1)
Original: qualifieddialogue-transfer-prior · gpt-5.6-sol · fable-5.1
A field-average emission/extinction ratio is not a universal cloud calibration
Mehandiratta finds a regional ratio3.67±.05 versus previous5.42±.05 and simulations favor cloud differences. These two means do not directly give a universal per-layer population width; geometry cancellation requires matched weighting.
Scientific value: Supports a hierarchical environment-dependent transfer with uncertainty.
Original passages (4)
Original: supporteddialogue-3d-inversion · gpt-5.6-sol
3D voxel counts are not independent polarized layer measurements
Extinction constrains where dust may lie. Zelko fits integrated intensity with extensive regularization; its reported two-temperature Cepheus demonstration is up to1σ. Neither product uniquely measures arbitrary layer polarized SEDs.
Scientific value: Requires radial covariance, unresolved components and selection functions in forward tests.
Original passages (2)
Original: supporteddialogue-mcbride-citation · fable-5.1 · grok-4.6 · gpt-5.6-sol
McBride and Sponseller are distinct papers with distinct inference setups
MNRAS5194370 is McBride,Bull&Hensley2023; Sponseller&Kogut is ApJ9368(2022). Their model complexity and estimators differ, so one result cannot substitute for the other.
Scientific value: Repairs the evidence chain behind claims of biased or unbiased recovery.
Original passages (4)
Original: qualifieddialogue-mcbride-theorem · grok-4.6 · fable-5.1
No-angle-mixing recovery is not an experiment-independent unbiased-r theorem
McBride reports generally unbiased CMBQ/U for tested simple configurations without simultaneous SED and angle variation. This does not prove unbiased r under arbitrary temperature distributions, instruments, higher-order SEDs, or estimators.
Scientific value: Promotes independent truth families and experiment-specific likelihood checks.
Original passages (2)
Original: supporteddialogue-spin-citation · fable-5.1 · gpt-5.6-sol
Spin-2 map moments and BB spectrum moments must be distinguished
The general spin-2 moment paper is A&A669A5(2023;2022preprint); Vacher2022A&A660A111 develops angular spectrum moment fitting. Real/imaginary rotation diagnostics must be relative to the zeroth spin moment.
Scientific value: Correct representation and prior-art attribution for implementation.
Original passages (4)
Original: incorrectdialogue-scalar-morphology · fable-5.1 · gpt-5.6-sol
Distinct scalar layer morphology is helpful, not a universal necessary condition
Noncollinear known polarization directions, spectral differences and external constraints can identify contributions even with proportional scalar intensity templates. The full Stokes/frequency design matrix matters.
Scientific value: Improves identifiability tests beyond scalar-map correlations.
Original passages (2)
Original: qualifieddialogue-fixed-mbb-column · grok-4.6 · gpt-5.6-sol
A nonlinear MBB can still have a conditional mixing-matrix column
At fixed spectral parameters, an MBB defines a frequency column. A spatially varying fitted MBB is not necessarily one globally separable sky template. Both the model parameters and spatial domain must be specified.
Scientific value: Makes rank-based comparisons precise without rejecting useful linearization.
Original passages (2)
Original: unsupporteddialogue-mask-gain · fable-5.1
Field selection is not automatically a pure gain
Masking can reduce foreground complexity but loses sky modes and changes variance, mode coupling and survey depth. Bias reduction need not improve total inference. Nearby dust dominance does not make all low-ℓ emission nearby.
Scientific value: Compare masks at fixed observing resources and report uncertainty with bias.
Original passages (2)
Original: incorrectdialogue-temperature-14k · fable-5.1 · grok-4.6
The 14K conversion of a15% band-ratio shift is only a local linearization
At19.6K dln(B353/B217)/dT=.01043797K^-1. Exact fixed-β ratios relative to19.6K are−15% at11.2269K and+15% at73.2409K. The paper’s cloud-ratio scatter is not a single pairwise finite shift.
Scientific value: Avoids a false exclusion of temperature models and supports nonlinear population simulation.
Original passages (2)
Original: qualifieddialogue-temperature-dominance · gpt-5.6-sol · fable-5.1
Planck release dependence prevents a universal temperature-dominance inference
Guillet finds temperature-only residual covariance compatible with SRoll2, while neither tested one-parameter variation model fits PR4. This is conditional evidence with a data-release discrepancy.
Scientific value: Treats β/T model choice as an unresolved empirical comparison.
Original passages (3)
Original: incorrectdialogue-reference-cancellation · fable-5.1
Reference-band depolarization does not guarantee small target-band residual
Two nearly canceling353GHz layers can have unequal150GHz scaling and leave substantial target-band polarization. A small reference total can make relative extrapolation errors large.
Scientific value: Include cancellation regimes in simulation and avoid dividing by an almost-zero total.
Original passages (1)
Original: incorrectdialogue-control-mbb · grok-4.6
A single HI cloud is not guaranteed to be a single polarized MBB
A velocity-cloud label can contain internal SED/field variation, unresolved beam structure, and missed gas phases. The control sample should be tested for excess relative to its model, not forced to a physically exact null.
Scientific value: Makes target/control null tests realistic and falsifiable.
Original passages (2)
Original: incorrectdialogue-rotation-sign · grok-4.6
HI angle alone does not predict the sign of frequency rotation
Interchanging which fixed-angle layer has the rising relative SED reverses the signed rotation. An angle proxy needs signed layer-to-SED association and amplitude information for that prediction.
Scientific value: Tests the joint association rather than only cloud geometry.
Original passages (1)
Original: incorrectdialogue-planck-limit · fable-5.1
Planck LR71 data-fit, simulation bound and confidence level are different quantities
G6 optimist uses0.975 as the largest-region correlation bound. PlanckXI Table5 gives LR71 data-fit0.989±.005 and E2E lower-limit0.991 at50≤ℓ≤160. The synthesis improves this distinction.
Scientific value: Maintains mask, statistic and simulation provenance for quantitative constraints.
Original passages (1)
Original: incorrectdialogue-r-sign · fable-5.1 · grok-4.6
Positive residual auto-power does not force every inferred-r bias upward
The cleaned spectrum shift includes2C(B,R)+C(R,R), not only the positive residual auto-term. Cleaning correlations, likelihood marginalization and parameter boundaries affect inferred r.
Scientific value: Prevents toy residual power being mislabeled a universal signed likelihood bias.
Original passages (2)
Original: qualifieddialogue-r-forecast · fable-5.1 · grok-4.6
One-pixel polarized residuals are not full BB or r forecasts
Spatially varying angle rotation couples E/B modes. A pixel residual fraction lacks a mask, angular covariance, transfer function, noise, delensing and likelihood. r-equivalent normalizations must be explicitly illustrative.
Scientific value: Separates mechanistic simulations from observational sensitivity forecasts.
Original passages (4)
Original: incorrectdialogue-single-band-beta · grok-4.6
One353GHz measurement cannot identify a dust spectral index even at fixed temperature
G6 line290 says a single fixed-temperature MBB is identified at353GHz. In a reference-normalized MBB the measurement determines only a dust Stokes amplitude (subject to other components); β is unconstrained at the pivot.
Scientific value: Provides a basic parameter-identifiability check.
Original passages (1)
Original: incorrectdialogue-normalized-derivative · fable-5.1
Reference-normalized temperature moments must subtract the pivot derivative
For gν∝Bν(T)/Bν0(T), ∂Tlngν=∂TlnBν−∂TlnBν0. G2 line185 omits the second term with a reference-amplitude convention. G6 critic and synthesis correctly distinguish the conventions.
Scientific value: Ensures the temperature perturbation is zero at its normalization band.
Original passages (1)
Original: qualifieddialogue-harmonic-factorization · gpt-5.6-sol · fable-5.1
Layer SEDs factor outside spectra only when spatially constant
C_l(ab)=Σg_i(a)g_j(b)C_l(ij) assumes fixed layer scalings over the region. Spatially varying g_i must remain inside the harmonic transform, or be expanded with additional fields.
Scientific value: Prevents an inconsistent covariance implementation.
Original passages (3)
Original: qualifieddialogue-hix-correlation · gpt-5.6-sol · fable-5.1
BKXVI correlation strength is not automatically explained dust power
The10–65% result concerns correlation of HI templates with BICEP/Keck maps including dust,CMB,noise. Turning it into a dust-power or r-bias fraction requires the paper’s estimator and component/noise treatment.
Scientific value: Calibrates what HI adds without converting unlike statistics.
Original passages (2)
Original: proposaldialogue-ancillary-frontier · gpt-5.6-sol
Measure when uncertain ancillary priors improve coverage and prediction
Use blinded independent truth models, realistic mock ancillary catalogues, fixed observing resources, held-out frequencies and ablations. Sweep missing layers, wrong associations, SED errors and angle errors to identify improvement and failure regimes.
Scientific value: A useful research direction; not an already demonstrated discovery or guaranteed novel method.
Original passages (6)
Original: incorrectSUM-01 · orchestrator summary
The summary inherits three incorrect bibliographic identities
The literature anchors give the wrong Ritacco journal locator, attribute McBride/Bull/Hensley to Hensley/Clark, and replace Delabrouille with Huffenberger in the 3D-model paper. These are real papers with incorrect bibliographic identities in the summary.
Scientific value: Correct attribution makes the useful scientific synthesis traceable.
Original passages (1)
04-summary/literature-anchors.md
Original: incorrectSUM-02 · orchestrator summary
The Pelgrims 60° selection was not a fabricated threshold
The summary presents the threshold as a citation-hygiene problem because it is absent from an abstract. The full paper explicitly uses a ≥60° misalignment cut and discusses an alternative ≥45° cut.
Scientific value: An abstract-only check can introduce a false correction; reproduce methods from the full text.
Original passages (1)
Original: qualifiedSUM-03 · orchestrator summary
Realized SED–angle covariance is not a population-correlation requirement
The exact phase derivative depends on a realized weighted complex covariance. That finite-sightline covariance can be nonzero even for independent draws of spectral parameters and orientations. Different spectra and physical angles also do not guarantee rotation, especially at 90° or when multiple pair terms cancel.
Scientific value: Use exact spin-2 diagnostics rather than inferring an unmeasured causal coupling.
Original passages (1)
Original: qualifiedSUM-04 · orchestrator summary
The known-angle condition number needs both spin-2 singularities
The summary repeats the useful two-amplitude idea but distinct physical angles are insufficient: separation by 90° gives antiparallel Stokes columns. The exact Euclidean condition number is sqrt[(1+|cos2Δψ|)/(1−|cos2Δψ|)], singular at 0° and 90°. One frequency suffices only for dust-only data with exactly two amplitudes and known noncollinear angles.
Scientific value: The audit derives the elementary linear algebra explicitly and supplies an interactive conditioning test; no priority claimed.
Original passages (1)
Original: unsupportedSUM-05 · orchestrator summary
Critic–optimist success in this corpus is not a general mode ranking
One prompt, one set of runs, shared later-stage context and an unblinded narrative comparison cannot show that drive is generally bad at correction or that critic–optimist is generally superior. The audit also finds corrections and regressions that the summary misses.
Scientific value: Retain the experiment as qualitative provenance; a performance claim requires repeated randomized runs and blinded claim scoring.
Original passages (2)
Original: qualifiedSUM-06 · orchestrator summary
Ideas unique within the model matrix largely have prior art
LOS optical-tomography motivation, spin moments, chromatic E/B and EB, and compression of 3D dust complexity have already been developed in papers. The repository supplies a useful synthesis but no new observed effect, benchmark result or validated r improvement.
Scientific value: The concrete extension is a controlled held-out comparison of tomography-conditioned, empirical and matched sham priors.
Original passages (1)