ResearchTasksModel ClassificationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSD200 zero-shot (holdout)MVProbe95.53Accuracy7May 25, 2026SD1k zero-shot (holdout)ProbeX50.5Accuracy (zero-shot)5Feb 26, 2026SD1k in-distributionDense38.2Accuracy5Feb 26, 2026SD200 in-distributionProbeX97.3Accuracy5Feb 26, 2026