List your company

Workflow Steering Attack on ASB-Bench (100 samples evaluation)

52TASRFLOWSTEER
35.3639.684448.32May 12, 2026

Evaluation results

MethodLinks
FLOWSTEERModel Family=GPT-4o, Planner-Executor Setup=G12026.055259
FLOWSTEERModel Family=GPT-4o, Planner-Executor Setup=G22026.055162
Direct Goal Injection (DGI)Model Family=Gemini-2.5, Planner-Executor Setup=G12026.055152
FLOWSTEERModel Family=Gemini-2.5, Planner-Executor Setup=G12026.055163
Direct Goal Injection (DGI)Model Family=DeepSeek, Planner-Executor Setup=G32026.055155
FLOWSTEERModel Family=DeepSeek, Planner-Executor Setup=G32026.055166
FLOWSTEERModel Family=GPT-4o, Planner-Executor Setup=G32026.053647