{"url":"/sota/domain-generalization-on-imagenet-a","task":{"name":"Domain Generalization","url":"/task/domain-generalization","note":null},"dataset":{"name":"ImageNet-A","url":"/dataset/imagenet-a"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"The idea of **Domain Generalization** is to learn from one or multiple training domains, to extract a domain-agnostic model which can be applied to an unseen domain\n\n\n<span class=\"description-source\">Source: [Diagram Image Retrieval using Sketch-Based Deep Learning and Transfer Learning ](https://arxiv.org/abs/2004.10780)</span>","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["Top-1 accuracy %","Number of params"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Top-1 accuracy %":"higher","Number of params":"lower"}},"counts":{"rows":39,"rows_with_code":38,"rows_with_paper_page":39,"rows_dated":39,"rows_using_additional_data":13},"rows":[{"rank_in_archive_order":1,"model":"Model soups (BASIC-L)","metrics":{"Top-1 accuracy %":"94.17"},"uses_additional_data":true,"paper_date":"2022-03-10","paper":"/paper/model-soups-averaging-weights-of-multiple","paper_url":"https://arxiv.org/abs/2203.05482v3","paper_title":"Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time","code":"https://github.com/mlfoundations/model-soups","n_code_links":6,"syntology":{"n_ran":5,"n_unverified":12,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"Model soups (ViT-G/14)","metrics":{"Top-1 accuracy %":"92.67"},"uses_additional_data":true,"paper_date":"2022-03-10","paper":"/paper/model-soups-averaging-weights-of-multiple","paper_url":"https://arxiv.org/abs/2203.05482v3","paper_title":"Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time","code":"https://github.com/mlfoundations/model-soups","n_code_links":6,"syntology":{"n_ran":5,"n_unverified":12,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"µ2Net+ (ViT-L/16)","metrics":{"Top-1 accuracy %":"84.53"},"uses_additional_data":true,"paper_date":"2022-09-15","paper":"/paper/a-continual-development-methodology-for-large","paper_url":"https://arxiv.org/abs/2209.07326v3","paper_title":"A Continual Development Methodology for Large-scale Multitask Dynamic ML Systems","code":"https://github.com/google-research/google-research/tree/master/muNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":4,"model":"CAR-FT (CLIP, ViT-L/14@336px)","metrics":{"Top-1 accuracy %":"81.5"},"uses_additional_data":true,"paper_date":"2022-11-29","paper":"/paper/context-aware-robust-fine-tuning","paper_url":"https://arxiv.org/abs/2211.16175v1","paper_title":"Context-Aware Robust Fine-Tuning","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":5,"model":"CAFormer-B36 (IN-21K, 384)","metrics":{"Number of params":"99M","Top-1 accuracy %":"79.5"},"uses_additional_data":true,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"MAE (ViT-H, 448)","metrics":{"Top-1 accuracy %":"76.7"},"uses_additional_data":false,"paper_date":"2021-11-11","paper":"/paper/masked-autoencoders-are-scalable-vision","paper_url":"https://arxiv.org/abs/2111.06377v2","paper_title":"Masked Autoencoders Are Scalable Vision Learners","code":"https://github.com/facebookresearch/mae","n_code_links":58,"syntology":{"n_ran":71,"n_unverified":66,"n_samples":137,"n_pointer_only_licence":73}},{"rank_in_archive_order":7,"model":"FAN-Hybrid-L(IN-21K, 384)","metrics":{"Top-1 accuracy %":"74.5"},"uses_additional_data":true,"paper_date":"2022-04-26","paper":"/paper/understanding-the-robustness-in-vision","paper_url":"https://arxiv.org/abs/2204.12451v4","paper_title":"Understanding The Robustness in Vision Transformers","code":"https://github.com/nvlabs/fan","n_code_links":2,"syntology":null},{"rank_in_archive_order":8,"model":"ConvFormer-B36 (IN-21K, 384)","metrics":{"Number of params":"100M","Top-1 accuracy %":"73.5"},"uses_additional_data":true,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"CAFormer-B36 (IN-21K)","metrics":{"Number of params":"99M","Top-1 accuracy %":"69.4"},"uses_additional_data":true,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"ConvNeXt-XL (Im21k, 384)","metrics":{"Top-1 accuracy %":"69.3"},"uses_additional_data":true,"paper_date":"2022-01-10","paper":"/paper/a-convnet-for-the-2020s","paper_url":"https://arxiv.org/abs/2201.03545v2","paper_title":"A ConvNet for the 2020s","code":"https://github.com/keras-team/keras/blob/master/keras/applications/convnext.py","n_code_links":54,"syntology":{"n_ran":54,"n_unverified":26,"n_samples":80,"n_pointer_only_licence":11}},{"rank_in_archive_order":11,"model":"MAE+DAT (ViT-H)","metrics":{"Top-1 accuracy %":"68.92"},"uses_additional_data":false,"paper_date":"2022-09-16","paper":"/paper/enhance-the-visual-representation-via","paper_url":"https://arxiv.org/abs/2209.07735v1","paper_title":"Enhance the Visual Representation via Discrete Adversarial Training","code":"https://github.com/alibaba/easyrobust/tree/main/examples/imageclassification/imagenet/dat","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":1,"n_samples":11,"n_pointer_only_licence":1}},{"rank_in_archive_order":12,"model":"ConvFormer-B36 (IN-21K)","metrics":{"Number of params":"100M","Top-1 accuracy %":"63.3"},"uses_additional_data":true,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":13,"model":"Pyramid Adversarial Training Improves ViT (Im21k)","metrics":{"Top-1 accuracy %":"62.44"},"uses_additional_data":true,"paper_date":"2021-11-30","paper":"/paper/pyramid-adversarial-training-improves-vit","paper_url":"https://arxiv.org/abs/2111.15121v2","paper_title":"Pyramid Adversarial Training Improves ViT Performance","code":"https://github.com/google-research/scenic/tree/main/scenic/projects/adversarialtraining","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"CAFormer-B36 (384)","metrics":{"Number of params":"99M","Top-1 accuracy %":"61.9"},"uses_additional_data":false,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":15,"model":"TransNeXt-Base (IN-1K supervised, 384)","metrics":{"Number of params":"89.7M","Top-1 accuracy %":"61.6"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/transnext-robust-foveal-visual-perception-for","paper_url":"https://arxiv.org/abs/2311.17132v3","paper_title":"TransNeXt: Robust Foveal Visual Perception for Vision Transformers","code":"https://github.com/Westlake-AI/openmixup","n_code_links":4,"syntology":{"n_ran":5,"n_unverified":0,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":16,"model":"TransNeXt-Small (IN-1K supervised, 384)","metrics":{"Number of params":"49.7M","Top-1 accuracy %":"58.3"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/transnext-robust-foveal-visual-perception-for","paper_url":"https://arxiv.org/abs/2311.17132v3","paper_title":"TransNeXt: Robust Foveal Visual Perception for Vision Transformers","code":"https://github.com/Westlake-AI/openmixup","n_code_links":4,"syntology":{"n_ran":5,"n_unverified":0,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"ConvFormer-B36 (384)","metrics":{"Number of params":"100M","Top-1 accuracy %":"55.3"},"uses_additional_data":false,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":18,"model":"SEER (RegNet10B)","metrics":{"Top-1 accuracy %":"52.7"},"uses_additional_data":true,"paper_date":"2022-02-16","paper":"/paper/vision-models-are-more-robust-and-fair-when","paper_url":"https://arxiv.org/abs/2202.08360v2","paper_title":"Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision","code":"https://github.com/facebookresearch/vissl","n_code_links":1,"syntology":null},{"rank_in_archive_order":19,"model":"TransNeXt-Base (IN-1K supervised, 224)","metrics":{"Number of params":"89.7M","Top-1 accuracy %":"50.6"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/transnext-robust-foveal-visual-perception-for","paper_url":"https://arxiv.org/abs/2311.17132v3","paper_title":"TransNeXt: Robust Foveal Visual Perception for Vision Transformers","code":"https://github.com/Westlake-AI/openmixup","n_code_links":4,"syntology":{"n_ran":5,"n_unverified":0,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":20,"model":"CAFormer-B36","metrics":{"Number of params":"99M","Top-1 accuracy %":"48.5"},"uses_additional_data":false,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":21,"model":"TransNeXt-Small (IN-1K supervised, 224)","metrics":{"Number of params":"49.7M","Top-1 accuracy %":"47.1"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/transnext-robust-foveal-visual-perception-for","paper_url":"https://arxiv.org/abs/2311.17132v3","paper_title":"TransNeXt: Robust Foveal Visual Perception for Vision Transformers","code":"https://github.com/Westlake-AI/openmixup","n_code_links":4,"syntology":{"n_ran":5,"n_unverified":0,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":22,"model":"FAN-L-Hybrid+STL","metrics":{"Top-1 accuracy %":"46.1"},"uses_additional_data":false,"paper_date":"2024-01-08","paper":"/paper/fully-attentional-networks-with-self-emerging-1","paper_url":"https://arxiv.org/abs/2401.03844v1","paper_title":"Fully Attentional Networks with Self-emerging Token Labeling","code":"https://github.com/NVlabs/STL","n_code_links":1,"syntology":null},{"rank_in_archive_order":23,"model":"ConvFormer-B36","metrics":{"Number of params":"100M","Top-1 accuracy %":"40.1"},"uses_additional_data":false,"paper_date":"2022-10-24","paper":"/paper/metaformer-baselines-for-vision","paper_url":"https://arxiv.org/abs/2210.13452v4","paper_title":"MetaFormer Baselines for Vision","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":8,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":24,"model":"Pyramid Adversarial Training Improves ViT (384x384)","metrics":{"Top-1 accuracy %":"36.41"},"uses_additional_data":false,"paper_date":"2021-11-30","paper":"/paper/pyramid-adversarial-training-improves-vit","paper_url":"https://arxiv.org/abs/2111.15121v2","paper_title":"Pyramid Adversarial Training Improves ViT Performance","code":"https://github.com/google-research/scenic/tree/main/scenic/projects/adversarialtraining","n_code_links":1,"syntology":null},{"rank_in_archive_order":25,"model":"Sequencer2D-L","metrics":{"Top-1 accuracy %":"35.5"},"uses_additional_data":false,"paper_date":"2022-05-04","paper":"/paper/sequencer-deep-lstm-for-image-classification","paper_url":"https://arxiv.org/abs/2205.01972v4","paper_title":"Sequencer: Deep LSTM for Image Classification","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":5,"syntology":{"n_ran":4,"n_unverified":5,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":26,"model":"Discrete Adversarial Distillation (ViT-B/224)","metrics":{"Top-1 accuracy %":"31.8 "},"uses_additional_data":false,"paper_date":"2023-11-02","paper":"/paper/distilling-out-of-distribution-robustness-1","paper_url":"https://arxiv.org/abs/2311.01441v2","paper_title":"Distilling Out-of-Distribution Robustness from Vision-Language Foundation Models","code":"https://github.com/lapisrocks/DiscreteAdversarialDistillation","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":27,"model":"Diffusion Classifier","metrics":{"Top-1 accuracy %":"30.2"},"uses_additional_data":false,"paper_date":"2023-03-28","paper":"/paper/your-diffusion-model-is-secretly-a-zero-shot","paper_url":"https://arxiv.org/abs/2303.16203v3","paper_title":"Your Diffusion Model is Secretly a Zero-Shot Classifier","code":"https://github.com/diffusion-classifier/diffusion-classifier","n_code_links":4,"syntology":{"n_ran":2,"n_unverified":0,"n_samples":2,"n_pointer_only_licence":1}},{"rank_in_archive_order":28,"model":"RVT-B*","metrics":{"Top-1 accuracy %":"28.5"},"uses_additional_data":false,"paper_date":"2021-05-17","paper":"/paper/rethinking-the-design-principles-of-robust","paper_url":"https://arxiv.org/abs/2105.07926v4","paper_title":"Towards Robust Vision Transformer","code":"https://github.com/alibaba/easyrobust","n_code_links":2,"syntology":null},{"rank_in_archive_order":29,"model":"RVT-S*","metrics":{"Top-1 accuracy %":"25.7"},"uses_additional_data":false,"paper_date":"2021-05-17","paper":"/paper/rethinking-the-design-principles-of-robust","paper_url":"https://arxiv.org/abs/2105.07926v4","paper_title":"Towards Robust Vision Transformer","code":"https://github.com/alibaba/easyrobust","n_code_links":2,"syntology":null},{"rank_in_archive_order":30,"model":"RVT-Ti*","metrics":{"Top-1 accuracy %":"14.4"},"uses_additional_data":false,"paper_date":"2021-05-17","paper":"/paper/rethinking-the-design-principles-of-robust","paper_url":"https://arxiv.org/abs/2105.07926v4","paper_title":"Towards Robust Vision Transformer","code":"https://github.com/alibaba/easyrobust","n_code_links":2,"syntology":null},{"rank_in_archive_order":31,"model":"GFNet-S","metrics":{"Top-1 accuracy %":"14.3"},"uses_additional_data":false,"paper_date":"2021-07-01","paper":"/paper/global-filter-networks-for-image","paper_url":"https://arxiv.org/abs/2107.00645v2","paper_title":"Global Filter Networks for Image Classification","code":"https://github.com/raoyongming/GFNet","n_code_links":4,"syntology":{"n_ran":6,"n_unverified":4,"n_samples":10,"n_pointer_only_licence":5}},{"rank_in_archive_order":32,"model":"CutMix+MoEx (ResNet-50)","metrics":{"Top-1 accuracy %":"8.4"},"uses_additional_data":false,"paper_date":"2020-02-25","paper":"/paper/on-feature-normalization-and-data","paper_url":"https://arxiv.org/abs/2002.11102v3","paper_title":"On Feature Normalization and Data Augmentation","code":"https://github.com/Boyiliee/MoEx","n_code_links":1,"syntology":{"n_ran":3,"n_unverified":11,"n_samples":14,"n_pointer_only_licence":0}},{"rank_in_archive_order":33,"model":"Discrete Adversarial Distillation (ResNet-50)","metrics":{"Top-1 accuracy %":"7.7"},"uses_additional_data":false,"paper_date":"2023-11-02","paper":"/paper/distilling-out-of-distribution-robustness-1","paper_url":"https://arxiv.org/abs/2311.01441v2","paper_title":"Distilling Out-of-Distribution Robustness from Vision-Language Foundation Models","code":"https://github.com/lapisrocks/DiscreteAdversarialDistillation","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":34,"model":"CutMix (ResNet-50)","metrics":{"Top-1 accuracy %":"7.3"},"uses_additional_data":false,"paper_date":"2019-05-13","paper":"/paper/cutmix-regularization-strategy-to-train","paper_url":"https://arxiv.org/abs/1905.04899v2","paper_title":"CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":30,"syntology":{"n_ran":17,"n_unverified":7,"n_samples":24,"n_pointer_only_licence":5}},{"rank_in_archive_order":35,"model":"Mixup (ResNet-50)","metrics":{"Top-1 accuracy %":"6.6"},"uses_additional_data":false,"paper_date":"2017-10-25","paper":"/paper/mixup-beyond-empirical-risk-minimization","paper_url":"http://arxiv.org/abs/1710.09412v2","paper_title":"mixup: Beyond Empirical Risk Minimization","code":"https://github.com/rwightman/pytorch-image-models","n_code_links":71,"syntology":{"n_ran":30,"n_unverified":17,"n_samples":47,"n_pointer_only_licence":15}},{"rank_in_archive_order":36,"model":"Cutout (ResNet-50)","metrics":{"Top-1 accuracy %":"4.4"},"uses_additional_data":false,"paper_date":"2017-08-15","paper":"/paper/improved-regularization-of-convolutional","paper_url":"http://arxiv.org/abs/1708.04552v2","paper_title":"Improved Regularization of Convolutional Neural Networks with Cutout","code":"https://github.com/albumentations-team/albumentations","n_code_links":28,"syntology":{"n_ran":21,"n_unverified":3,"n_samples":24,"n_pointer_only_licence":5}},{"rank_in_archive_order":37,"model":"ResNet-50 (300 Epochs)","metrics":{"Top-1 accuracy %":"4.2"},"uses_additional_data":false,"paper_date":"2015-12-10","paper":"/paper/deep-residual-learning-for-image-recognition","paper_url":"http://arxiv.org/abs/1512.03385v1","paper_title":"Deep Residual Learning for Image Recognition","code":"https://github.com/tensorflow/models/tree/master/research/deeplab","n_code_links":484,"syntology":{"n_ran":230,"n_unverified":147,"n_samples":377,"n_pointer_only_licence":187}},{"rank_in_archive_order":38,"model":"Stylized ImageNet (ResNet-50)","metrics":{"Top-1 accuracy %":"2.3"},"uses_additional_data":true,"paper_date":"2018-11-29","paper":"/paper/imagenet-trained-cnns-are-biased-towards","paper_url":"https://arxiv.org/abs/1811.12231v3","paper_title":"ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness","code":"https://github.com/rgeirhos/texture-vs-shape","n_code_links":7,"syntology":{"n_ran":0,"n_unverified":6,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":39,"model":"ResNet-50","metrics":{"Top-1 accuracy %":"0"},"uses_additional_data":false,"paper_date":"2019-07-16","paper":"/paper/natural-adversarial-examples","paper_url":"https://arxiv.org/abs/1907.07174v4","paper_title":"Natural Adversarial Examples","code":"https://github.com/hendrycks/natural-adv-examples","n_code_links":3,"syntology":{"n_ran":1,"n_unverified":3,"n_samples":4,"n_pointer_only_licence":0}}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,264 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9581,"papers_checked":6264,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":3316},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":29,"rows_with_any_sample_ran":20,"distinct_papers_with_graph_line":17,"distinct_papers_with_any_sample_ran":15,"samples_over_distinct_papers":{"n_ran":460,"n_unverified":312,"n_samples":772,"n_pointer_only_licence":303,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":481,"n_unverified":352,"n_samples":833,"n_pointer_only_licence":303,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}