Distractor Injection Attacks on Large Reasoning Models: Characterization and Defense

Open in new window