{"url":"/sota/out-of-distribution-detection-on-imagenet-1k-8","task":{"name":"Out-of-Distribution Detection","url":"/task/out-of-distribution-detection","note":null},"dataset":{"name":"ImageNet-1k vs SUN","url":"/dataset/imagenet-1k-vs-sun"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"Detect out-of-distribution or anomalous examples.","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":["FPR95","AUROC"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"FPR95":null,"AUROC":null}},"counts":{"rows":22,"rows_with_code":21,"rows_with_paper_page":22,"rows_dated":22,"rows_using_additional_data":2},"rows":[{"rank_in_archive_order":1,"model":"CMA(ViT-B/16, NegLabel)","metrics":{"AUROC":"96.36","FPR95":"16.84"},"uses_additional_data":false,"paper_date":"2025-03-24","paper":"/paper/enhanced-ood-detection-through-cross-modal","paper_url":"https://arxiv.org/abs/2503.18817v1","paper_title":"Enhanced OoD Detection through Cross-Modal Alignment of Multi-Modal Representations","code":"https://github.com/ma-kjh/CMA-OoDD","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"LINe (ResNet50)","metrics":{"AUROC":"95.26","FPR95":"19.48"},"uses_additional_data":false,"paper_date":"2023-03-24","paper":"/paper/line-out-of-distribution-detection-by","paper_url":"https://arxiv.org/abs/2303.13995v1","paper_title":"LINe: Out-of-Distribution Detection by Leveraging Important Neurons","code":"https://github.com/yonghyun-ahn/line-out-of-distribution-detection-by-leveraging-important-neurons","n_code_links":1,"syntology":{"n_ran":14,"n_unverified":6,"n_samples":20,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"NNGuide (RegNet)","metrics":{"AUROC":"94.43","FPR95":"21.58"},"uses_additional_data":true,"paper_date":"2023-09-26","paper":"/paper/nearest-neighbor-guidance-for-out-of-1","paper_url":"https://arxiv.org/abs/2309.14888v1","paper_title":"Nearest Neighbor Guidance for Out-of-Distribution Detection","code":"https://github.com/jingkang50/openood","n_code_links":2,"syntology":{"n_ran":6,"n_unverified":2,"n_samples":8,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"SCALE (ResNet50)","metrics":{"AUROC":"95.02","FPR95":"23.27"},"uses_additional_data":false,"paper_date":"2023-09-30","paper":"/paper/scaling-for-training-time-and-post-hoc-out-of","paper_url":"https://arxiv.org/abs/2310.00227v1","paper_title":"Scaling for Training Time and Post-hoc Out-of-distribution Detection Enhancement","code":"https://github.com/kai422/scale","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"DICE + ReAct (ResNet-50)","metrics":{"AUROC":"93.94","FPR95":"25.45"},"uses_additional_data":false,"paper_date":"2021-11-18","paper":"/paper/on-the-effectiveness-of-sparsification-for","paper_url":"https://arxiv.org/abs/2111.09805v2","paper_title":"DICE: Leveraging Sparsification for Out-of-Distribution Detection","code":"https://github.com/deeplearning-wisc/dice","n_code_links":1,"syntology":{"n_ran":14,"n_unverified":5,"n_samples":19,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"ASH-S (ResNet-50)","metrics":{"AUROC":"94.02","FPR95":"27.98"},"uses_additional_data":false,"paper_date":"2022-09-20","paper":"/paper/extremely-simple-activation-shaping-for-out","paper_url":"https://arxiv.org/abs/2209.09858v2","paper_title":"Extremely Simple Activation Shaping for Out-of-Distribution Detection","code":"https://github.com/andrijazz/ash","n_code_links":2,"syntology":{"n_ran":2,"n_unverified":0,"n_samples":2,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"MCM (CLIP-L)","metrics":{"AUROC":"94.14","FPR95":"29.00"},"uses_additional_data":false,"paper_date":"2022-11-24","paper":"/paper/delving-into-out-of-distribution-detection","paper_url":"https://arxiv.org/abs/2211.13445v1","paper_title":"Delving into Out-of-Distribution Detection with Vision-Language Representations","code":"https://github.com/deeplearning-wisc/mcm","n_code_links":2,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":1}},{"rank_in_archive_order":8,"model":"RankFeat (ResNetv2-101)","metrics":{"AUROC":"94.07","FPR95":"29.27"},"uses_additional_data":false,"paper_date":"2022-09-18","paper":"/paper/rankfeat-rank-1-feature-removal-for-out-of","paper_url":"https://arxiv.org/abs/2209.08590v1","paper_title":"RankFeat: Rank-1 Feature Removal for Out-of-distribution Detection","code":"https://github.com/kingjamessong/rankfeat","n_code_links":1,"syntology":null},{"rank_in_archive_order":9,"model":"DICE (ResNet-50)","metrics":{"AUROC":"90.83","FPR95":"35.15"},"uses_additional_data":false,"paper_date":"2021-11-18","paper":"/paper/on-the-effectiveness-of-sparsification-for","paper_url":"https://arxiv.org/abs/2111.09805v2","paper_title":"DICE: Leveraging Sparsification for Out-of-Distribution Detection","code":"https://github.com/deeplearning-wisc/dice","n_code_links":1,"syntology":{"n_ran":14,"n_unverified":5,"n_samples":19,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"MCM (CLIP-B)","metrics":{"AUROC":"92.57","FPR95":"37.59"},"uses_additional_data":false,"paper_date":"2022-11-24","paper":"/paper/delving-into-out-of-distribution-detection","paper_url":"https://arxiv.org/abs/2211.13445v1","paper_title":"Delving into Out-of-Distribution Detection with Vision-Language Representations","code":"https://github.com/deeplearning-wisc/mcm","n_code_links":2,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":1}},{"rank_in_archive_order":11,"model":"MOS (BiT-S-R101x1)","metrics":{"AUROC":"92.01","FPR95":"40.63"},"uses_additional_data":true,"paper_date":"2021-05-05","paper":"/paper/mos-towards-scaling-out-of-distribution","paper_url":"https://arxiv.org/abs/2105.01879v1","paper_title":"MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Space","code":"https://github.com/deeplearning-wisc/large_scale_ood","n_code_links":5,"syntology":{"n_ran":3,"n_unverified":1,"n_samples":4,"n_pointer_only_licence":1}},{"rank_in_archive_order":12,"model":"NPOS","metrics":{"AUROC":"90.44","FPR95":"43.77"},"uses_additional_data":false,"paper_date":"2023-03-06","paper":"/paper/non-parametric-outlier-synthesis","paper_url":"https://arxiv.org/abs/2303.02966v1","paper_title":"Non-Parametric Outlier Synthesis","code":"https://github.com/deeplearning-wisc/npos","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":1}},{"rank_in_archive_order":13,"model":"GradNorm (ResNetv2-101)","metrics":{"FPR95":"46.48"},"uses_additional_data":false,"paper_date":"2021-10-01","paper":"/paper/on-the-importance-of-gradients-for-detecting","paper_url":"https://arxiv.org/abs/2110.00218v2","paper_title":"On the Importance of Gradients for Detecting Distributional Shifts in the Wild","code":"https://github.com/deeplearning-wisc/gradnorm_ood","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":0,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":14,"model":"ReAct (ResNet-50)","metrics":{"AUROC":"88.16","FPR95":"47.69"},"uses_additional_data":false,"paper_date":"2021-11-24","paper":"/paper/react-out-of-distribution-detection-with-1","paper_url":"https://arxiv.org/abs/2111.12797v1","paper_title":"ReAct: Out-of-distribution Detection With Rectified Activations","code":"https://github.com/deeplearning-wisc/react","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":1,"n_samples":2,"n_pointer_only_licence":2}},{"rank_in_archive_order":15,"model":"KNN (ResNet-50 SupCon)","metrics":{"AUROC":"88.40","FPR95":"48.91"},"uses_additional_data":false,"paper_date":"2022-04-13","paper":"/paper/out-of-distribution-detection-with-deep","paper_url":"https://arxiv.org/abs/2204.06507v3","paper_title":"Out-of-Distribution Detection with Deep Nearest Neighbors","code":"https://github.com/deeplearning-wisc/knn-ood","n_code_links":3,"syntology":{"n_ran":30,"n_unverified":8,"n_samples":38,"n_pointer_only_licence":38}},{"rank_in_archive_order":16,"model":"ODIN+UMAP (ResNet-50)","metrics":{"AUROC":"86.92","FPR95":"49.69"},"uses_additional_data":false,"paper_date":"2023-06-06","paper":"/paper/unleashing-mask-explore-the-intrinsic-out-of","paper_url":"https://arxiv.org/abs/2306.03715v1","paper_title":"Unleashing Mask: Explore the Intrinsic Out-of-Distribution Detection Capability","code":"https://github.com/tmlr-group/unleashing-mask","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":5,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"SHE","metrics":{"FPR95":"54.19"},"uses_additional_data":false,"paper_date":"2023-01-21","paper":"/paper/out-of-distribution-detection-based-on-in","paper_url":"https://openreview.net/forum?id=KkazG4lgKL","paper_title":"Out-of-Distribution Detection based on In-Distribution Data Patterns Memorization with Modern Hopfield Energy","code":"https://github.com/zjs975584714/she_ood_detection","n_code_links":1,"syntology":null},{"rank_in_archive_order":18,"model":"DML","metrics":{"FPR95":"57.40"},"uses_additional_data":false,"paper_date":"2023-01-01","paper":"/paper/decoupling-maxlogit-for-out-of-distribution","paper_url":"http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Decoupling_MaxLogit_for_Out-of-Distribution_Detection_CVPR_2023_paper.html","paper_title":"Decoupling MaxLogit for Out-of-Distribution Detection","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":19,"model":"KNN (ResNet-50)","metrics":{"AUROC":"80.10","FPR95":"69.53"},"uses_additional_data":false,"paper_date":"2022-04-13","paper":"/paper/out-of-distribution-detection-with-deep","paper_url":"https://arxiv.org/abs/2204.06507v3","paper_title":"Out-of-Distribution Detection with Deep Nearest Neighbors","code":"https://github.com/deeplearning-wisc/knn-ood","n_code_links":3,"syntology":{"n_ran":30,"n_unverified":8,"n_samples":38,"n_pointer_only_licence":38}},{"rank_in_archive_order":20,"model":"RP+GradNorm","metrics":{"FPR95":"73.53"},"uses_additional_data":false,"paper_date":"2023-04-25","paper":"/paper/detecting-out-of-distribution-data-through-in","paper_url":"https://openreview.net/forum?id=charggEv8v&referrer=%5Bthe%20profile%20of%20Tongliang%20Liu%5D(%2Fprofile%3Fid%3D~Tongliang_Liu1)","paper_title":"Detecting Out-of-distribution Data through In-distribution Class Prior","code":"https://github.com/tmlr-group/class_prior","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"DOE","metrics":{"AUROC":"76.26","FPR95":"80.94"},"uses_additional_data":false,"paper_date":"2023-03-09","paper":"/paper/out-of-distribution-detection-with-implicit","paper_url":"https://arxiv.org/abs/2303.05033v1","paper_title":"Out-of-distribution Detection with Implicit Outlier Transformation","code":"https://github.com/qizhouwang/doe","n_code_links":1,"syntology":null},{"rank_in_archive_order":22,"model":"MOOD","metrics":{"AUROC":"89.8"},"uses_additional_data":false,"paper_date":"2023-02-06","paper":"/paper/rethinking-out-of-distribution-ood-detection","paper_url":"https://arxiv.org/abs/2302.02615v2","paper_title":"Rethinking Out-of-distribution (OOD) Detection: Masked Image Modeling is All You Need","code":"https://github.com/dvlab-research/mood","n_code_links":2,"syntology":null}],"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 6,885 of the 9,623 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":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+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":15,"rows_with_any_sample_ran":15,"distinct_papers_with_graph_line":12,"distinct_papers_with_any_sample_ran":12,"samples_over_distinct_papers":{"n_ran":82,"n_unverified":29,"n_samples":111,"n_pointer_only_licence":43,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":130,"n_unverified":43,"n_samples":173,"n_pointer_only_licence":82,"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"}}}