List your company

Safety Evaluation on CocoNot

0.613Safety ScoreGRAPH ROUTER
0.15020.270350.39050.51065Feb 5, 2026

Evaluation results

MethodLinks
GRAPH ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Benign)2026.020.613
LLM ROUTERMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.575
LLM ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.569
LLM ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Benign)2026.020.568
LLM ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.549
LLM ROUTERMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.541
GRAPH ROUTERMalicious Setting=M4-RL, Mitigation Strategy=supervisor-free mitigation2026.020.495
GRAPH ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.492
GRAPH ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-free mitigation2026.020.49
GRAPH ROUTERMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.485
GRAPH ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.484
GRAPH ROUTERMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.46
LLM ROUTERMalicious Setting=M4-RL, Mitigation Strategy=supervisor-free mitigation2026.020.429
LLM ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-free mitigation2026.020.424
LLM ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Malicious)2026.020.41
TEXT DEBATEMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.373
LLM ROUTERMalicious Setting=M4-RL, Mitigation Strategy=None (Malicious)2026.020.368
TEXT DEBATEMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-free mitigation2026.020.365
TEXT DEBATEMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.365
TEXT DEBATEMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.333
TEXT DEBATEMalicious Setting=M4-RL, Mitigation Strategy=supervisor-free mitigation2026.020.332
TEXT FEEDBACKMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-free mitigation2026.020.319
TEXT DEBATEMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.307
TEXT FEEDBACKMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.304
TEXT DEBATEMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Benign)2026.020.298
TEXT FEEDBACKMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Benign)2026.020.296
TEXT FEEDBACKMalicious Setting=M2-Activation Steering, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.291
TEXT FEEDBACKMalicious Setting=M4-RL, Mitigation Strategy=None (Malicious)2026.020.254
TEXT DEBATEMalicious Setting=M4-RL, Mitigation Strategy=None (Malicious)2026.020.242
TEXT FEEDBACKMalicious Setting=M4-RL, Mitigation Strategy=supervisor-free mitigation2026.020.242
TEXT FEEDBACKMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S1)2026.020.233
TEXT FEEDBACKMalicious Setting=M4-RL, Mitigation Strategy=supervisor-based mitigation (S2)2026.020.232
TEXT DEBATEMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Malicious)2026.020.2
TEXT FEEDBACKMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Malicious)2026.020.199
GRAPH ROUTERMalicious Setting=M4-RL, Mitigation Strategy=None (Malicious)2026.020.184
GRAPH ROUTERMalicious Setting=M2-Activation Steering, Mitigation Strategy=None (Malicious)2026.020.168