Pith.

Every reviewed paper

Each paper is listed with what Pith's machine review of it concluded.

498,816

papers reviewed

live · read 16 Nov 2026, 03:23 UTC

Narrow the list

Page 7. Ordered by the date each paper was submitted, newest first.

  1. Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    The passes disagreed. Reader: holds up. Skeptic: holds up only with conditions. Referee: major revision.

    A transparent multi-epoch ALMA study whose month-scale variability and ~10^3 R_S causality limit for NGC 985's compact millimeter component hold up well, with one unresolved QA-execution question a referee should press on.

    4 major objections, the first at §2.1, §4.2 · 8 minor

  2. The paper:

    Proton Spin Fluctuation Decomposition

    Physics · arXiv:2611.17032

    Tomoya Uji

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    A genuine new formal framework for proton spin fluctuations, but the key derivative identities connecting double GTMDs to OAM-weighted distributions are asserted rather than derived, which is the load-bearing gap.

    the referee pass did not run

  3. The paper:

    Your Vision Foundation Model Is Secretly an EEG Foundation Model

    Computer Science · arXiv:2611.17031

    Guoan Wang, Feng Liu

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Frozen ViT beats all EEG foundation models on 8 datasets—the evidence is strong and the method is clean. The cross-fitting concern is real but the paper acknowledges it and the …

    1 major objection at §3.3, Eq. (7) · 6 minor

  4. The paper:

    Similar but Not Redundant: Embedding-Gated Refinement for High-Granularity Topic Models

    Computer Science · arXiv:2611.17030

    Daniel Denley, Robert L. Nicol, Mike J. Chantler

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A real methodological contribution with large, consistent effects, held back by an oracle whose ground truth is the same conservative LLM that makes the merges.

    2 major objections, the first at §4.4, Appendix A.3 (Prompt 4), Table 7 · 6 minor

  5. The paper:

    Contractual Structure and Obvious Manipulation in Many-to-Many Matching

    Economics · arXiv:2611.17029

    R. Pablo Arribillaga, Beatriz Millan, Eliana Pepa Risma

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Solid extension of NOM to many-to-many matching with contracts; the efficiency result (Theorem 1) is clean and the negative conditions are well-motivated, but the positive results (Theorems 5 and 6) are narrower than they first appear.

    the referee pass did not run

  6. The paper:

    ALMA quiet Sun observations compared with a simulated MURaM-ChE chromosphere

    Physics · arXiv:2611.17028

    Maria Loukitcheva, Sami K. Solanki, Damien Przybylski and 2 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Honest model validation that lands on the right side of cautious, but the single-snapshot premise is load-bearing and the 'closest agreement' claim is softer than the abstract suggests.

    3 major objections, the first at Abstract and §3.3.2, Tables 3-4 · 5 minor

  7. The paper:

    Spin to win: enhancing the quantum Otto cycle with the circular Unruh effect

    Physics · arXiv:2611.17027

    Leo J. A. Parry, Dimitris Moustos, Benito A. Ju\'arez-Aubry and 2 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    The passes disagreed. Reader: holds up. Skeptic: holds up only with conditions. Referee: major revision.

    3 major objections, the first at §3.1, §3.2.3, §5 · 8 minor

  8. Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Strongly convex ATPD results are solid and match lower bounds; the general convex O(1/N^2) rate is conditional on an unproven uniform boundedness assumption that the authors acknowledge but do not resolve.

    1 major objection at Theorem 5.1 and Remark 5.1 · 4 minor

  9. The paper:

    Spin-Boson Hamiltonians as Generators of Spin-Dressed Bogoliubov Dynamics

    Physics · arXiv:2611.17025

    Nathan Metraud

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A mathematically rigorous dynamical definition of spin-boson Hamiltonians that unifies regular and renormalized models; the central uniqueness theorem has a load-bearing gap at the singular end of its stated range.

    2 major objections, the first at Theorem C(ii) / Theorem 3.15(ii) and Remark 3.24 · 5 minor

  10. Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A clean, honestly-reported GMM case study for player archetypes whose abstract overclaims 'objective' hybrid-role measurement, though the authors' own discussion mostly corrects it.

    3 major objections, the first at Section 3.1.4 and Abstract · 5 minor

  11. The paper:

    Sedimentation of flexible fibers through an ordered array of pillars

    Physics · arXiv:2611.17023

    Ursy Makanga, Camille Duprat, Blaise Delmotte

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Clear numerical phenomenology of fiber modes in a pillar array; the theoretical model is a sketch, not a predictor, and the paper's own stated assumptions limit its reach.

    the referee pass did not run

  12. The paper:

    Large diameter for (Δ + 2)-colouring reconfiguration

    Mathematics · arXiv:2611.17022

    Wouter Cames van Batenburg

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up. The paper's explicit constructions genuinely kill the n+mu conjecture from CCC24; I checked the delicate counting in Theorem 3 and it holds.

    This review says the paper's mathematics is wrong. Pith did not check that against the paper's source.

  13. The paper:

    Entanglement detection from moments of the realigned state: application to continuous-variable quantum systems

    Physics · arXiv:2611.17021

    Serge Deside, Vojt\v{e}ch Kala, Nicolas J. Cerf

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Solid CV entanglement witness based on realigned moments, experimentally accessible, but provably weaker than PPT and cannot detect bound entanglement.

    3 major objections, the first at §V.B.2, Eq. (118) and Fig. 6 · 7 minor

  14. The paper:

    Maximizing LLM Utility While Managing LLM Fallibility

    Computer Science · arXiv:2611.17020

    Ekta U. Samani, Christopher G. Atkeson

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up. A competent, honest review whose contribution is its taxonomy of LLM roles, reliance dimensions, reliability questions, and mapped classical methods; the transfer-from-classical-methods claim is plausible but empirically thin.

    3 major objections, the first at Section 4 and Section 6 · 5 minor

  15. The paper:

    Temporal Regularized Learning: Self-Supervised Learning Local in Space and Time

    Computer Science · arXiv:2611.17019

    Davide Wiest

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A clean, honest, reproducible local learning rule with real but small gains — the sharpest unresolved question is the boundary oracle the class-chunked stream hands the network.

    2 major objections, the first at Section 4, Eq. (2), Section 2.4 · 4 minor

  16. The paper:

    Exploring Socially Acceptable 3D Feedback Design for Inclusive Small-Group Conversation

    Computer Science · arXiv:2611.17018

    Qianjie Wei, Yifan Li, Zhen Bai

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Solid design-space exploration with honest methods, but the headline claim about reduced discomfort understates how thin the inferential support is outside Scenario 1.

    3 major objections, the first at §5.2.2 and §10 (conclusion, b297) · 6 minor

  17. The paper:

    Continuum space hierarchical Cannings process: II. Genealogies and ancestral web

    Mathematics · arXiv:2611.17017

    A. Greven, F. den Hollander

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A real construction of continuum-space genealogies, an ancestral web, and a L\'evy web with strong duality, but the load-bearing block-resampling extension is deferred to in-preparation companions; referee it, with the companions demanded.

    5 major objections, the first at §3.3.2, Proposition 3.12 · 8 minor

  18. The paper:

    Control Variates to reduce the variance of the empirical Moment Generating Function: an extreme value perspective

    Mathematics · arXiv:2611.17016

    Louison Bocquet-Nouaille, J\'er\^ome Morio, Benjamin Bobbia and 1 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Does not hold up as stated. Interesting qualitative insight, but the displayed formulas (4) and (5) are wrong as written due to mishandling of slowly varying functions.

    This review says the paper's mathematics is wrong. Pith did not check that against the paper's source.

    5 major objections, the first at Appendix A.3, Eq. (4) and surrounding derivation · 8 minor

  19. The paper:

    AI for Biologists: The Knowledge Boundaries of Foundation Models in the Age of Agentic Systems

    Biology · arXiv:2611.17015

    Tian Cai, Andrew Kleist, Jaimin Patel and 5 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A genuinely useful, balanced review of where biological foundation models fail, wrapped around a five-element framework that is asserted rather than demonstrated, and whose proposed boundary statements would not have caught the paper's own failure examples.

    2 major objections, the first at Abstract; §2.1.2; §4.3; §6.2 · 4 minor

  20. The paper:

    Poincaré polynomials of nilmanifolds associated with path graphs

    Mathematics · arXiv:2611.17014

    Marco Aldi, Samuel Bevins, Eleftherios Chatzitheodoridis and 3 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. The paper delivers explicit closed-form Poincaré polynomials for path-graph nilmanifolds, which is a genuine advance within a niche area. The main result is sound, but the proof of the load-bearing …

    3 major objections, the first at Section 3, Lemma 3.2 (Eq. (7)) · 2 minor

  21. The paper:

    DiffSSMEye: Conditional High-Frequency Eye Movement Synthesis with Structured State-Space Diffusion

    Computer Science · arXiv:2611.17013

    Kamrul Hasan, Mehedi Hasan Raju, Oleg V. Komogortsev

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Solid spatial results, but the biometric claim is a self-consistency check—authors disclose this, but the abstract doesn't.

    3 major objections, the first at §5.3, Eq. (20); §3.5.2, Eq. (12); §3.3, Eq. (3) · 7 minor

  22. Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. The empirical core is real and the critique is sound, but the paper's load-bearing assumption—that the corpus is exactly what the EC would release—is unverifiable and funded by a party with aligned interests.

    5 major objections, the first at §5, §6, Acknowledgments · 7 minor

  23. The paper:

    Random Exploration is Random

    Biology · arXiv:2611.17011

    Yingkang Luo, Hua-Dong Xiong, Hanbo Xie and 5 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A careful negative-result study with a stronger title than its evidence: flexible models improve fit but don't close the short- versus long-horizon predictability gap, which is a meaningful result even if 'truly substantially random' overstates it.

    3 major objections, the first at §2.5, Fig. 3 · 9 minor

  24. Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up. Solid ALICE measurement of Lambda(1520) suppression at 5.02 TeV; the 6.8-sigma central-to-peripheral suppression is the real news, and the model comparison is honest. …

    3 major objections, the first at §1, pT-integrated yield extraction · 10 minor

  25. The paper:

    Modelling redshift-space distortions with tracer-centred split densities

    Physics · arXiv:2611.17009

    Aritra K. Gon, Yan-Chuan Cai, Mark Neyrinck and 3 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions.

    4 major objections, the first at Section 5, Fig. 11 and Fig. B4 · 6 minor

  26. The paper:

    Sticky Manifold Piecewise Deterministic Monte Carlo for Bayesian Model Averaging

    Statistics · arXiv:2611.17008

    Hirofumi Shiba, Luke Hardcastle, Kengo Kamatani

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A principled, modular framework for trans-dimensional PDMP that generalises sticky PDMPs to curved submanifolds; the core invariance proof is solid, but the velocity compatibility condition is a genuine practical bottleneck that the paper acknowledges but does not resolve.

    4 major objections, the first at Section 3.1, Eq. (3.2) · 7 minor

  27. The paper:

    Skill Adapter: Personalized Modification and Online Routing of General-Purpose Skills

    Computer Science · arXiv:2611.17007

    Qijia Zhuang, Yanyi Gou, Wenbin Ma and 1 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Solid engineering paper on skill personalization and routing for LLM agents; the routing reward's reference set is underspecified, but the empirical gains are large and the ablation is thorough.

    4 major objections, the first at §Reinforcement Learning, Eq. (14) · 8 minor

  28. The paper:

    Infinite spin superparticles

    Physics · arXiv:2611.17006

    Eugenia Boffo, Martin Cederwall

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Supersymmetric infinite spin particles from an action principle, with honest limitations.

    4 major objections, the first at §4.1.3, footnote 3 · 5 minor

  29. The paper:

    An Integrated Architecture for Data-Efficient Real-Time Neural Imitation of Nonlinear Time-Optimal Model Predictive Control

    Systems · arXiv:2611.17005

    Maarten De Baecke, Flavia Sofia Acerbo, Andrea Giusti and 1 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. SIFS is a genuinely new sampling idea for NNMPC, but the central data-efficiency claim is undercut by a one-sided stopping rule in the baseline comparison.

    4 major objections, the first at Sec. VI.B (Fig. 5b); abstract; Sec. VIII · 8 minor

  30. The paper:

    Real log canonical tresholds and real log canonical tresholds with signs

    Mathematics · arXiv:2611.17004

    Th{\'e}o Jaudon (IRMAR, UR)

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    First closed-form signed real lct for plane curves, but the main theorem leans on an unpublished preprint for the load-bearing pole structure.

    the referee pass did not run

  31. The paper:

    Maturity2: Assessing the Maturity of Crypto Agility Maturity Models

    Computer Science · arXiv:2611.17003

    Leonie Wolf, Robin Leander Schr\"oder, Sebastian Schinzel and 1 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    The passes disagreed. Reader: holds up. Skeptic: holds up only with conditions. Referee: major revision.

    Systematic evaluation of three industry crypto agility models against design principles; the central finding holds up, but the stress-test concern about applying a full maturity-model lens to artifacts that never …

    3 major objections, the first at §3.1, §5 · 8 minor

  32. The paper:

    Hausdorff measures of planar Brownian level sets

    Mathematics · arXiv:2611.17002

    Yingyu Zhao, Yunhao Zhao

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Resolves Taylor's 40-year-old gauge question for planar Brownian level sets with a sharp iff condition, and the proofs hold up; the open constants and the exact spectrum are honestly acknowledged as open ends, not holes.

    the referee pass did not run

  33. The paper:

    Reflection Zeros Set the Scattering Invariant of Finite Majorana Wires

    Physics · arXiv:2611.17001

    Alex Krasnok

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    The passes disagreed. Reader: holds up. Skeptic: agrees with the reader. Referee: major revision.

    A clean counting rule for Q in finite Majorana wires, backed by solid algebra and extensive numerics. The paper is honest about its limitations and deserves a serious referee.

    4 major objections, the first at Derivation of Eq. (1) and wide-band limit · 8 minor

  34. The paper:

    Macroscopic L∞ bounds for the homogeneous Landau–Coulomb equation

    Mathematics · arXiv:2611.17000

    Lingda Xu

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A technically impressive a priori estimate that resolves the open problem under the assumption of smooth, rapidly decaying solutions; the gap to general initial data is real but clearly scoped.

    2 major objections, the first at §1.1 and Abstract · 6 minor

  35. The paper:

    Dynamical mean-field theory of latching Potts networks

    Biology · arXiv:2611.16999

    Anindita Basu, Carlo Orientale Caputo, Fernando Lucas Metz and 1 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. DMFT for Potts latching networks: technically solid, central claim about S>1 necessity is well-supported at α=0 but not yet proven at finite load.

    4 major objections, the first at §IV C, Fig. 7, and §V · 6 minor

  36. Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up. A careful NEGF study of the spinterface model that finds a structural zero and a clean linear/quadratic hierarchy under non-collinear leads; the central claims are analytically grounded and numerically verified, …

    2 major objections, the first at Appendix B · 6 minor

  37. The paper:

    Entire sig-jump-diffusions, their expected signature and applications to score-based generative modeling

    Mathematics · arXiv:2611.16995

    Christa Cuchiero, Francesca Primavera, Sara Svaluto-Ferro

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    The passes disagreed. Reader: holds up, with conditions. Skeptic: does not hold up. Referee: major revision.

    The paper extends signature SDEs to jump processes and shows that the expected signature solves a linear ODE; the generative modeling application is a plausible sketch but rests on an unverified entire approximation and has no numerical validation.

    4 major objections, the first at Section 4.3, Theorem 4.9 · 6 minor

  38. Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Real contribution is the ellipse-projection interval that handles the d_0 = 0 corner; the load-bearing fragility is Assumption 1 with an uninformative benchmark, and the negative-control calibration that would patch it is missing from the England demo.

    the referee pass did not run

  39. The paper:

    It Was Better Not to Know: Digital Recognition, Family Authority, and Postpartum Mental Health in Bangladesh

    Computer Science · arXiv:2611.16993

    Shajnin Haque Aishy, Sabbir Bin Abdul Latif, Md. Golam Zel A Husna and 1 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up. A well-executed qualitative study that names a real structural gap—self-recognition without family recognition—and earns its design implications. The sample limits are honestly stated and the central claim holds up.

    4 major objections, the first at §3.1, Table 1 · 7 minor

  40. The paper:

    Media Trading with Censored Auction Prices: A Stochastic Control and Reinforcement Learning Approach

    Finance · arXiv:2611.16992

    Yun-Ithry Gamrani, Huy\^en Pham

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A clean formulation of a practical problem, with honest limitations; the central estimator bound has a known gap with the experiments, but the paper is transparent about it and the framework is useful.

    4 major objections, the first at §4.2, Lemma 4.1 and Remark 4.1 · 12 minor

  41. The paper:

    AnalogLayoutWorld: An Open Benchmark for Agentic Analog Layout Design

    Computer Science · arXiv:2611.16991

    NovaSilicon Team

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    A solid, well-executed benchmark for analog layout agents; the engineering quality metric is the main soft spot, but the paper acknowledges it and the central claim holds up.

    the referee pass did not run

  42. The paper:

    How to arrange four points in the plane

    Mathematics · arXiv:2611.16990

    Senping Luo

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    The passes disagreed. Reader: holds up. Skeptic: agrees with the reader. Referee: major revision.

    Complete classification of four-point minimizers for general Lennard-Jones potentials; rigorous and self-contained, but limited to N=4.

    5 major objections, the first at §2, Lemma 2.2 · 10 minor

  43. The paper:

    NΩ Dibaryon near Unitarity from Physical-Point Lattice QCD

    Physics · arXiv:2611.16989

    Liang Zhang, Yan Lyu, Sinya Aoki and 4 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    The passes disagreed. Reader: holds up. Skeptic: holds up only with conditions. Referee: major revision.

    First physical-point lattice QCD result for NΩ binding, well executed but the single-channel assumption is the load-bearing soft spot.

    3 major objections, the first at HAL QCD method, sentence b30 · 6 minor

  44. The paper:

    Unconstrained ℤ2 gauge theory on a triangular lattice: deconfined fermions in a vison crystal and superfluidity

    Physics · arXiv:2611.16988

    Welberth Kennedy, Jo\~ao C. In\'acio, Fakher F. Assaad

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. The vison crystal phase and the first-order transition are solid; the deconfinement claim is asserted but not demonstrated.

    3 major objections, the first at Sec. IV C, Fig. 9 · 9 minor

  45. The paper:

    A lift of Baxter's cyclic summation identity with applications to quantum topology

    Mathematics · arXiv:2611.16987

    Calvin McPhail-Snyder

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. Phase-free lift of Baxter's cyclic sum is real and well-supported; the branch-choice concern in the stress test is real but the paper handles it better than the note suggests.

    3 major objections, the first at §2.2, Theorem 2.6, Eq. (2.11) · 10 minor

  46. The paper:

    Primer: An AI Framework for Instructor-Centered Educational Simulation Creation

    Computer Science · arXiv:2611.16986

    Ethan Mollick, Lilach Mollick, Ben Przystanski and 3 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions.

    2 major objections, the first at Abstract; Section 'Voice, Pedagogy, and Risks'; Appendix B · 5 minor

  47. The paper:

    ElasticDino: Dense Semantic Features Using Elastic Deformations

    Computer Science · arXiv:2611.16984

    Ulysse Mizrahi, Sean Zaretzky, Dov Danon and 1 more

    Pith says

    Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded

    Holds up. ElasticDino is a clean, well-executed paper that introduces a deformation-based upsampling method for ViT features, with strong empirical results and a clear architectural insight.

    4 major objections, the first at Section 3.2, Eq. (1) · 6 minor