List your company

OS Assistance on Safe-OS system-sabotage and environment-dependent

100TSRbaseline
-4235077Jun 14, 2026

Evaluation results

MethodLinks
baselineAccess Control=LLM-AC, Core LLM=Gemini 2.5 Flash2026.0610010
FragFuseAccess Control=LLM-AC, Core LLM=Gemini 2.5 Flash2026.0684.464
direct queryingAccess Control=None, Core LLM=GPT-5.12026.0684—
direct queryingAccess Control=None, Core LLM=Gemini 2.5 Flash2026.0682—
direct queryingAccess Control=None, Core LLM=Average2026.0682—
FragFuseAccess Control=LLM-AC, Core LLM=Average2026.0681.480.7
FragFuseAccess Control=AGrail, Core LLM=GPT-5.12026.0681.396
FragFuseAccess Control=AGrail, Core LLM=Gemini 2.5 Flash2026.0680.994
FragFuseAccess Control=LLM-AC, Core LLM=GPT-4o2026.0680.582
direct queryingAccess Control=None, Core LLM=GPT-4o2026.0680—
FragFuseAccess Control=LLM-AC, Core LLM=GPT-5.12026.0679.296
FragFuseAccess Control=AGrail, Core LLM=Average2026.0677.692.7
FragFuseAccess Control=AGrail, Core LLM=GPT-4o2026.0670.588
baselineAccess Control=AGrail, Core LLM=GPT-4o2026.064010
baselineAccess Control=LLM-AC, Core LLM=Average2026.0633.33.3
baselineAccess Control=AGrail, Core LLM=Average2026.0621.419.3
baselineAccess Control=AGrail, Core LLM=GPT-5.12026.0614.328
baselineAccess Control=AGrail, Core LLM=Gemini 2.5 Flash2026.061020
baselineAccess Control=LLM-AC, Core LLM=GPT-4o2026.0600
baselineAccess Control=LLM-AC, Core LLM=GPT-5.12026.0600