ResearchTasksGeneralizationGeneralizationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedLLaMA Original Languages 3.1-8BLLaMA-3.1-8B-Instruct80.37Generalization Score8May 19, 2026LLaMA Evaluation Expanded Languages 3.1-8BDeltaMoE69.37Overall Score8May 19, 2026GenModular Gradient Surgery35.7Gen Score8Feb 26, 2026Single scenario unseen tasksFedRouter58.3ROUGE-16Apr 2, 2026Synthetic DatasetAdverISF (Two-Stage)0.251R^2 (10%)5Feb 26, 2026