ResearchBenchmarksInstruction Hierarchy Robustness on TensorTrust sys-userInstruction Hierarchy Robustness on TensorTrust sys-userFollow94ScoreGPT-5-Mini-R85.6887.849092.16Mar 11, 2026Evaluation resultsMethodMethodLinksScoreGPT-5-Mini-R2026.0394GPT-5-Mini2026.0386