ResearchDatasetsReal AgentsFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsSteering ValidationReal Agents LLMAgent, Claude Haiku 4.5 N-scaling Phase 4 (val)83.7Haiku Judge Accuracy4
Steering ValidationReal Agents LLMAgent, Claude Haiku 4.5 N-scaling Phase 4 (val)83.7Haiku Judge Accuracy4