Inverting the Shield: Systematically Generating Safety Tests from Policy Specifications · 相关技术
相关技术
rsasafetyiOSAPIANNARGCamouflaged object detectionEFIENAEffort Metric AttentionFOLfirst-order logicForHISLLMlarge language modelsLMMMITMaliNATNarrative Abstraction BenchmarkODEORMOWLOnline surgical phase recognitionPPRPROBESAFESEMSSEStraight-Through EstimatorTEEUARUCTdivide-and-conquer partitioningevallanguage model