Pith. sign in

REVIEW 3 major objections 5 minor 18 references

Beyond Attention: Signed Integrated Gradients Attribution in a BiomeGPT-Style Microbiome Transformer

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Signed attribution of the IBD–healthy margin replaces unsigned attention in feature-tokenized transformer explanations.

desk verdict The baseline derivation is correct and worth stealing; the empirical case for the biological claims is thin and the straight-line path assumption needs work. read the letter →

arxiv 2608.06486 v1 pith:CZ6GBDBM submitted 2026-08-06 cs.LG

classification cs.LG
keywords IntegratedGradientsfeature-tokenizedtransformermicrobiomeattentionexplanationIBDclassificationsignedattributionbaselineselectionBiomeGPT
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that the widely used [CLS]-attention weights cannot explain decisions of feature-tokenized transformers such as the microbiome model BiomeGPT: attention is nonnegative and operates on fused species-abundance tokens, so a token can receive high attention while pushing the prediction away from the class it seems to support. The author proposes to replace attention with Integrated Gradients applied to the disease-versus-health logit margin $F = z_{\mathrm{IBD}} - z_{\mathrm{Healthy}}$, using a token-fusion baseline $T' = S + A_0$ that keeps the species identity fixed and interpolates only the abundance embedding from its learned zero state. Under additive token fusion this makes fused-token attribution equivalent to abundance-source attribution and yields signed values: positive supports IBD, negative supports healthy. If correct, the approach gives feature-tokenized transformers a principled, signed explanation that distinguishes disease-driving from protective evidence, and it motivates second-order Integrated Hessians for species-abundance interaction analysis.

What carries the argument

The load-bearing object is the source-derived baseline $T' = S + A_0$, with $A_0 = \mathrm{MLP}_{AB}(0)$ the learned zero-abundance embedding. Integrated Gradients of the margin $F(S,A) = z_{\mathrm{IBD}} - z_{\mathrm{Healthy}}$ are computed along the interpolation $S_\alpha = S$, $A_\alpha = A_0 + \alpha(A - A_0)$; the identity-map fusion makes $\partial F/\partial S = \partial F/\partial A = \partial F/\partial T$ along the path, so the displacement factor $(A - A_0)$ alone distinguishes abundance-source attribution, and fused-token IG collapses to abundance IG. The species embedding is never removed; it stays in the forward pass and conditions the gradients, while the reported $G_A = \int_0^1 (\partial F/\partial A)(S, A_\alpha)\,d\alpha$ offers a non-complete sensitivity diagnostic that is not weighted by abundance displacement.

What would settle it

A concrete test: on the same trained classifier, compute abundance attributions with the source-derived baseline, with an all-zero fused-token baseline, and with an integration path that runs through the learned abundance encoder rather than straight-line in embedding space; if the sign assignments for the top-ranked species or the aggregate true-positive/true-negative mean signs are not stable across baselines and paths, the claimed directional evidence is an artifact of the chosen baseline rather than a property of the model.

Watch

Extended reading notes

Core claim

The central claim is that signed attribution of the IBD–healthy margin recovers directional species-abundance evidence that unsigned attention cannot express. In the author's construction, the baseline is derived from the token sources rather than set to zero: with $T = S + A$ and $A_0 = \mathrm{MLP}_{AB}(0)$, the baseline is $T' = S + A_0$, so the path moves abundance from its zero state to its observed value while the species embedding remains fixed. Because both sources enter the fusion additively with identity Jacobians, the path-averaged gradients of $F$ with respect to $S$ and $A$ coincide, and the fused-token IG equals the abundance-source IG; species attribution is zero by construction yet species identity conditions every gradient along the path. Empirically the author reports that across 7,382 samples true-positive samples carry positive mean attribution and true-negative samples negative mean attribution, and that several of the highest-[CLS]-attention species are among the strongest health-supporting signals rather than IBD drivers. The paper also introduces path-averaged abundance sensitivity $G_A$ and abundance-attribution response profiles as diagnostics.

Load-bearing premise

The argument rests on treating the learned zero-abundance embedding $A_0$ as a meaningful 'absence of signal' baseline and on assuming that straight-line interpolation in abundance-embedding space passes through regions where the model's gradients remain semantically interpretable, so a failure of either assumption would strip the signed attributions of biological meaning.

Editorial extensions

If this is right

  • Attention rankings should not be read as disease-driver rankings in feature-tokenized models, because the paper's results show top-attention species can be health-supporting.
  • The signed margin attribution gives each species-abundance observation a direction relative to the IBD–healthy decision, enabling lists of IBD-supporting versus health-supporting taxa from a single trained model.
  • Path-averaged abundance sensitivity $G_A$ can surface low-abundance species with high per-unit influence that full IG ranks low and attention ignores entirely.
  • Abundance-attribution response profiles show non-monotonic and context-dependent species effects, motivating segmented or path-resolved IG as the next diagnostic.
  • The baseline construction transfers to any additively fused feature-tokenized transformer, where the fixed identity embedding is preserved and only the value embedding is interpolated.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the additive-fusion equivalence is generic, then every additively fused feature-tokenized model inherits the same baseline rule: keep the feature's learned zero-value embedding fixed while preserving the feature identity, making the method portable beyond microbiome data.
  • The observation that high-attention species are often health-supporting suggests that attention-based biomarker extraction in other disease classifiers may systematically invert rankings; this is testable by rerunning the same signed-margin analysis on other fine-tuned feature-tokenized models.
  • The use of embedding-space interpolation is the delicate step; a tokenizer-aware path through the abundance encoder would give a direct robustness check and may change sign assignments for rare species.
  • The proposed second-order Integrated Hessian score $\Gamma^A_{ij}$ could be computed per species pair as a community-interaction map, offering a model-based alternative to correlation networks for candidate microbial interactions.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The manuscript proposes a source-derived Integrated Gradients baseline, T' = S + A0, for feature-tokenized transformers with additive token fusion, where S is a fixed species embedding and A is a sample-dependent abundance embedding. It proves that, under additive fusion, fused-token IG with this baseline is equivalent to abundance-source IG, with zero species attribution by construction. The method is then applied to a reimplemented BiomeGPT-style microbiome transformer for IBD versus healthy classification, and the paper reports signed attribution differences from [CLS] attention, path-averaged sensitivity diagnostics, and abundance-attribution response profiles. The central argument is that signed IG separates disease-supporting from health-supporting evidence where unsigned attention weights cannot, and the paper proposes second-order Integrated Hessians as a future extension for species-abundance interactions.

Significance. If the empirical and path-interpretation issues are resolved, the paper makes a clean and useful contribution: the derivation in Section 3.4 is correct and self-contained, and the source-derived baseline generalizes naturally to other additive feature-tokenized architectures such as FT-Transformer. The availability of code and the detailed reimplementation description are strengths. However, the empirical demonstration is currently preliminary: aggregate statistics are reported without uncertainty, the tested model is a reimplementation rather than the released BiomeGPT, and the biological interpretation of the attribution signs depends on an unvalidated straight-line path in embedding space. These issues limit the strength of the claims that the method 'uncovers species-abundance directional relationships' and 'separates pathogenic from protective microbial signals.'

major comments (3)
  1. [Section 3.4 and Section 6] The central interpretative claim, that IGA measures the effect of moving species abundance from zero to its observed value, requires the linear path A0 + alpha*(A - A0) to traverse abundance-embedding points that correspond to valid abundance variation. The paper itself cites UDIG (Ref. 13) for the concern that straight-line interpolation is poorly matched to tokenized models and explicitly defers a tokenizer-aware path to future work in Section 6, yet no experiment checks whether intermediate path points decode to valid abundance inputs, whether A0 is in-distribution, or whether the attribution signs are stable under an alternative path. Without such diagnostics, the signed rankings in Figures 4 and 6 may be artifacts of off-manifold interpolation rather than evidence about actual abundance effects, so the biological interpretation is not yet established.
  2. [Section 5] The quantitative evidence for the central empirical claim is thin. The only aggregate attribution numbers are the overall mean (-0.0011), the true-positive mean (0.0027), and the true-negative mean (-0.0045), all reported without standard deviations, confidence intervals, or significance tests, and the baseline margin F(S, A0) is never reported. Completeness only relates the total attribution sum to F(x) - F(x'), so these aggregate means carry little per-species information. The claimed 'clear contrast' between attention and IG, and the directional species rankings in Figure 4, require per-species distributions, effect sizes, and preferably a statistical comparison. Please report the baseline margin, per-species and per-sample attribution distributions, and appropriate uncertainty estimates.
  3. [Section 4 and Abstract/Conclusion] The experiments use 'a BiomeGPT-style microbiome transformer encoder' rather than the actual released BiomeGPT model, but the abstract and conclusion attribute the empirical findings to 'BiomeGPT' without qualification. The manuscript does not state whether the reimplementation matches the released BiomeGPT tokenizer, architecture, pretraining corpus, or pretrained weights, nor whether any external validation was performed. Because the empirical results are the only support for the biological directional claims, please specify exactly which components are shared with BiomeGPT and which are new, and adjust the abstract and conclusion claims so they refer to the reimplementation unless the released model is actually used.
minor comments (5)
  1. [Section 3.4] The sentence 'preserves the fixed identity feature coordinate component and interpolate only the component encoding the observed value' has a grammatical subject-verb disagreement; it should read '...and interpolates only the component encoding the observed value.'
  2. [Section 4] The all-zero baseline is written as 'T' = 0 in R^512', but the fused input is a sequence of 1663 tokens each of dimension 512; please clarify whether the baseline is a per-token vector or a full sequence tensor.
  3. [Section 5] Please define 'True Positive' and 'True Negative' explicitly as correctly classified IBD samples and correctly classified healthy samples, respectively, since these terms could otherwise be confused with clinical ground truth.
  4. [Section 3.1] Please clarify whether abundance bin 0 is an actual discrete input bin in the tokenizer vocabulary or a special out-of-vocabulary input; the definition of A0 = MLP_AB(0) depends on this distinction.
  5. [Section 2] The claim that prior applications of IG to feature-tokenized architectures 'have not been explicitly derived or documented' is strong; please soften it to 'to the best of our knowledge' after a more thorough literature search, or cite the specific prior work that was checked.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the baseline and equivalence are derived from the model's tokenization equation, not from the target result or fitted data.

full rationale

The paper's central derivation in Section 3.4 is a mathematical identity, not a circular prediction. The fused-token baseline T' = S + A0 is defined from the model's additive tokenization T = S + A and the learned zero-abundance embedding A0 = MLPAB(0); it does not depend on the attribution target F or on any fitted attribution quantity. The claimed equivalence IGT = (T-T') * integral(dF/dT) = (A-A0) * integral(dF/dA) = IGA follows directly from the chain rule under additive fusion (dT/dS = dT/dA = I), and the paper states the zero species attribution as a consequence of holding species identity fixed rather than as an empirical discovery. The sign semantics (positive supports IBD, negative supports healthy) are explicitly tied to the chosen logit contrast F = z_IBD - z_Healthy, so the observed sign consistency in Section 5 is a restatement of the target definition, not an independent prediction. No parameter is fitted to the data to define the baseline, and no prediction is forced by construction: the empirical comparisons to [CLS] attention, the G_A sensitivity diagnostic, and the abundance-response profiles are computed from the model and are not assumed by the derivation. The related-work discussion cites UDIG to acknowledge that straight-line paths may be poorly matched to tokenized models, and Section 6 explicitly disclaims biological causality and defers tokenizer-aware paths; these are correctness or generalization limitations, not circularity. There are no load-bearing self-citations: the author cites external work for IG, Captum, UDIG, Integrated Hessians, and BiomeGPT, with no uniqueness theorem imported from the author's own prior work. Overall the derivation is self-contained and does not reduce to its inputs.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The central derivation rests on the additive-fusion assumption and on the semantic validity of the zero-abundance baseline. The analysis introduces two hand-chosen parameters: the IG integration discretization and the bottom-5% abundance cutoff. No new physical or conceptual entities are postulated.

free parameters (2)
  • IG integration discretization
    Number of integration steps and alpha grid used by Captum are not reported; they affect numerical completeness and, weakly, attribution values.
  • Bottom-5% abundance cutoff = 5%
    Hand-chosen exclusion for the Figure 4 rankings; post hoc and can change which species appear in top lists.
assumptions (6)
  • domain assumption Feature tokens are additively fused as T = S + A (species embedding plus abundance embedding).
    Stated in Section 3.1 Eq (1); the entire source-derived baseline and equivalence result depend on additive fusion.
  • domain assumption A0 = MLP_AB(0) is a meaningful zero-abundance baseline representing absence of the species.
    Introduced in Section 3.4; if this embedding is out-of-distribution or not semantically 'absence', the integration path and attributions are not interpretable.
  • standard math The trained model is smooth and differentiable enough for path integration of gradients.
    Required by Integrated Gradients; assumed for the transformer and MLP and standard in the attribution literature.
  • standard math Integrated Gradients completeness holds up to numerical error.
    Invoked in Section 3.3 as the link between attributions and the IBD-healthy margin; depends on accurate integration.
  • domain assumption The reimplemented 'BiomeGPT-style' model is a faithful proxy for BiomeGPT.
    Section 4 reimplements the encoder rather than using the original model; generalization of the empirical demonstration to actual BiomeGPT depends on this.
  • domain assumption Validation accuracy around 0.93 indicates the model learned an IBD-discriminative signal suitable for attribution.
    Section 4 uses this to justify analyzing attributions; it does not establish external or biological validity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Beyond Attention: Signed Integrated Gradients Attribution in a BiomeGPT-Style Microbiome Transformer." pith.science (2026). https://pith.science/paper/CZ6GBDBM

@misc{pith2026260806486,
  author       = {Pith},
  title        = {Pith review of: Beyond Attention: Signed Integrated Gradients Attribution in a BiomeGPT-Style Microbiome Transformer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CZ6GBDBM}},
  note         = {Machine review of arXiv:2608.06486}
}
read the original abstract

In a feature-tokenized transformer (arXiv:2106.11959) such as BiomeGPT (doi:10.64898/2026.01.05.697599), each input token is built by fusing a fixed identity with a sample-specific measurement: a fixed species and a variable abundance, T = S + A. To interpret downstream classification in such models, prior work inspects the attention weights of the special [CLS] token (arXiv:2106.11959, arXiv:1810.04805, BiomeGPT) to rank sample tokens by importance. These weights have two critical limitations: they are nonnegative, so they cannot separate disease-supporting from health-supporting evidence (arXiv:2201.12114), and they act after token fusion, obscuring how the input sources S and A each affect the output. To address this we use Integrated Gradients (arXiv:1703.01365), a signed, fusion-aware attribution method, and propose a source-derived baseline T' = S + A_0 for feature-tokenized models such as BiomeGPT, which preserves species identity as a fixed biological coordinate while isolating the effect of abundance variation. Applied to a disease-versus-health decision margin, it yields polarity that explicitly separates pathogenic from protective microbial signals. We show that this gradient-based approach uncovers species-abundance directional relationships and sensitivity diagnostics entirely obscured by unsigned [CLS] attention weights. We further recommend second-order Integrated Hessians (arXiv:2002.04138) to expose microbiome community interaction rules: how a perturbation in one member alters the model's sensitivity to another, and which other species drive ambiguous cases toward disease or health at a given abundance level. This provides a principled approach to explainability in BiomeGPT that generalizes to other smooth and differentiable feature-tokenized transformers. Code is available at https://github.com/nohren/token-source-attribution

Figures

Figures reproduced from arXiv: 2608.06486 by the authors.

Figure 1
Figure 1. The species embedding remains fixed, while the abundance embedding varies across [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. The yellow dotted line represents the integration steps from [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Comparison of [CLS] attention weights and Integrated Gradients for species–abundance tokens on a single inference: one true positive sample (CSM79HM9_P, IBDMDB cohort) classified as IBD with pIBD = 0.999. Tokens are annotated (species, binned abundance). In the IG panels, red denotes positive attribution, increasing the IBD–Healthy logit margin; green denotes negative attribution, supporting the healthy class; uncol… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Top species rankings for IBD classification. Left: species receiving the highest [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Top species ranked by path-averaged abundance sensitivity [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Abundance–attribution response profiles for four species. Each point represents one [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 7 canonical work pages

  1. [1]

    Towards better understanding of gradient-based attribution methods for deep neural networks

    Marco Ancona, Enea Ceolini, Cengiz Öztireli, and Markus Gross. Towards better understanding of gradient-based attribution methods for deep neural networks. InInternational Conference on Learning Representations, 2018. URLhttps://arxiv.org/abs/1711.06104. arXiv:1711.06104

  2. [2]

    McIver, Kelsey N

    Aitor Blanco-Míguez, Francesco Beghini, Fabio Cumbo, Lauren J. McIver, Kelsey N. Thompson, Moreno Zolfo, Paolo Manghi, Leonard Dubois, Kun D. Huang, Andrew Maltez Thomas, William A. Nickols, Gianmarco Piccinno, Elisa Piperni, Michal Punˇcocháˇr, Mireia Valles-Colomer, Adrian Tett, Francesca Gior- dano, Richard Davies, Jonathan Wolf, Sarah E. Berry, Tim D....

  3. [3]

    scGPT: toward building a foundation model for single-cell multi-omics using generative AI.Nature Meth- ods, 21:1470–1480, 2024

    Haotian Cui, Chloe Wang, Hassaan Maan, Kuan Pang, Fengning Luo, Nan Duan, and Bo Wang. scGPT: toward building a foundation model for single-cell multi-omics using generative AI.Nature Meth- ods, 21:1470–1480, 2024. doi: 10.1038/s41592-024-02201-0. URL https://doi.org/10.1038/ s41592-024-02201-0

  4. [4]

    Bert: Pre-training of deep bidi- rectional transformers for language understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidi- rectional transformers for language understanding. InProceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers), pages 4171–4186, 2019

  5. [5]

    Revisiting deep learning models for tabular data

    Yury Gorishniy, Ivan Rubachev, Valentin Khrulkov, and Artem Babenko. Revisiting deep learning models for tabular data. InAdvances in Neural Information Processing Systems (NeurIPS), vol- ume 34, pages 18932–18943, 2021. URL https://proceedings.neurips.cc/paper/2021/hash/ 9d86d83f925f2149e9edb0ac3b49229c-Abstract.html. 14

  6. [6]

    Sarthak Jain and Byron C. Wallace. Attention is not explanation. InProceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics, pages 3543–3556, 2019. URLhttps://arxiv.org/abs/1902.10186. arXiv:1902.10186

  7. [7]

    Explaining Explanations: Axiomatic Feature Interactions for Deep Networks

    Joseph D. Janizek, Pascal Sturmfels, and Su-In Lee. Explaining explanations: Axiomatic feature interac- tions for deep networks.arXiv preprint arXiv:2002.04138, 2020. doi: 10.48550/arXiv.2002.04138. URL https://arxiv.org/abs/2002.04138. arXiv:2002.04138

  8. [8]

    Captum: A unified and generic model interpretability library for pytorch, 2020

    Narine Kokhlikyan, Vivek Miglani, Miguel Martin, Edward Wang, Bilal Alsallakh, Jonathan Reynolds, Alexander Melnikov, Natalia Kliushkina, Carlos Araya, Siqi Yan, and Orion Reblitz-Richardson. Captum: A unified and generic model interpretability library for pytorch, 2020. URL https://arxiv.org/abs/ 2009.07896. arXiv:2009.07896

Show all 18 references
  1. [9]

    Rethinking attention- model explainability through faithfulness violation test

    Yibing Liu, Haoliang Li, Yangyang Guo, Chenqi Kong, Jing Li, and Shiqi Wang. Rethinking attention- model explainability through faithfulness violation test. InProceedings of the 39th International Conference on Machine Learning, pages 13858–13871. PMLR, 2022

  2. [10]

    Medearis, Siyao Zhu, and Ali R

    Nicholas A. Medearis, Siyao Zhu, and Ali R. Zomorrodi. Biomegpt: A foundation model for the human gut microbiome.bioRxiv, 2026. doi: 10.64898/2026.01.05.697599. URL https://doi.org/10.64898/ 2026.01.05.697599

  3. [11]

    Dowd, Curtis Huttenhower, Martin Morgan, Nicola Segata, and Levi Waldron

    Edoardo Pasolli, Lucas Schiffer, Paolo Manghi, Audrey Renson, Valerie Obenchain, Duy Tin Truong, Francesco Beghini, Faizan Malik, Marcel Ramos, Jennifer B. Dowd, Curtis Huttenhower, Martin Morgan, Nicola Segata, and Levi Waldron. Accessible, curated metagenomic data through Ex...

  4. [12]

    David, and Xiaoli Fern

    Quintin Pope, Rohan Varma, Christine Tataru, Maude M. David, and Xiaoli Fern. Learning a deep language model for microbiomes: The power of large scale unlabeled microbiome data.PLOS Computational Biology, 21(5):e1011353, 2025. doi: 10.1371/journal.pcbi.1011353. URL https://doi...

  5. [13]

    Uniform discretized integrated gradients: An effective attribution based method for explaining large language models, 2024

    Swarnava Sinha Roy and Ayan Kundu. Uniform discretized integrated gradients: An effective attribution based method for explaining large language models, 2024. URLhttps://arxiv.org/abs/2412.03886. arXiv:2412.03886

  6. [14]

    Axiomatic attribution for deep networks

    Mukund Sundararajan, Ankur Taly, and Qiqi Yan. Axiomatic attribution for deep networks. InInternational Conference on Machine Learning, pages 3319–3328, 2017. URL https://arxiv.org/abs/1703. 01365. arXiv:1703.01365

  7. [15]

    Theodoris, Ling Xiao, Anant Chopra, Mark D

    Christina V . Theodoris, Ling Xiao, Anant Chopra, Mark D. Chaffin, Zeina R. Al Sayed, Matthew C. Hill, Helene Mantineo, Elizabeth M. Brydon, Zexian Zeng, X. Shirley Liu, and Patrick T. Ellinor. Transfer learning enables predictions in network biology.Nature, 618(7965):616–624,...

  8. [16]

    Attention is not not explanation

    Sarah Wiegreffe and Yuval Pinter. Attention is not not explanation. InProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing, pages 11–20, 2019. URL https://arxiv.org/ abs/1908.04626. arXiv:1908.04626

  9. [17]

    scBERT as a large-scale pretrained deep language model for cell type annotation of single-cell RNA-seq data.Nature Machine Intelligence, 4:852–866, 2022

    Fan Yang, Wenchuan Wang, Fang Wang, Yuan Fang, Duyu Tang, Junzhou Huang, Hui Lu, and Jianhua Yao. scBERT as a large-scale pretrained deep language model for cell type annotation of single-cell RNA-seq data.Nature Machine Intelligence, 4:852–866, 2022. doi: 10.1038/s42256-022-0...

  10. [2023]

    URLhttps://doi.org/10.1038/s41587-023-01688-w

    doi: 10.1038/s41587-023-01688-w. URLhttps://doi.org/10.1038/s41587-023-01688-w

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.