List your company

Safety Detection on BeaverTails (held-out)

76.7AUROCMultiLayer-Linear
73.16474.0827575.918May 18, 2026

Evaluation results

MethodLinks
MultiLayer-LinearProtocol=Leave-one-benchmark-out (LOBO), Aggregation=Mean across nine model backbones2026.0576.7
Geometry-LiteProtocol=Leave-one-benchmark-out (LOBO), Aggregation=Mean across nine model backbones2026.0576.3
Best single-layer probeProtocol=Leave-one-benchmark-out (LOBO), Aggregation=Mean across nine model backbones2026.0575.4
TaT-Disp-LSTMProtocol=Leave-one-benchmark-out (LOBO), Aggregation=Mean across nine model backbones2026.0574.2
MultiLayer-DIMProtocol=Leave-one-benchmark-out (LOBO), Aggregation=Mean across nine model backbones2026.0573.3