ResearchTasksReward Hacking DetectionReward Hacking DetectionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedSchool of Reward Hacks and Stanford Alpaca (test)activation-based monitoring system100F1 Score18Mar 5, 2026Anthropic’s Auditing GameSPT96.6Acc (Poisoned)6Feb 26, 2026
School of Reward Hacks and Stanford Alpaca (test)activation-based monitoring system100F1 Score18Mar 5, 2026