List your company

General Language Understanding on MMLU (Utility)

73.5Utility ScoreBase
71.73272.19172.6573.109Jun 8, 2026

Evaluation results

MethodLinks
Base2026.0673.5
AdvGRPO (ST)Attacker=Qwen2.5-7B2026.0673.3
AdvGRPO (ST)Attacker=Qwen2.5-14B2026.0673.1
Self-RedTeamAttacker=Qwen2.5-7B2026.0671.9
AdvGameAttacker=Qwen2.5-7B2026.0671.8