List your company

Multi-label Content Safety Classification on Synthetic Dataset

0.768F1 ScoreMarginal Probability
0.44040.525450.61050.69555Nov 27, 2025

Evaluation results

MethodLinks
Marginal ProbabilityModel=LLaMA Guard 32025.110.7680.906
Marginal ProbabilityModel=LLaMA 3.1 8B Instruct2025.110.7380.934
Conditional ProbabilityModel=LLaMA 3.1 8B Instruct2025.110.7040.876
Greedy GenerationModel=LLaMA Guard 32025.110.701—
Conditional ProbabilityModel=LLaMA Guard 32025.110.7010.777
Joint ProbabilityModel=LLaMA Guard 32025.110.70.777
Probability UncertaintyModel=LLaMA Guard 32025.110.698—
Entropy UncertaintyModel=LLaMA Guard 32025.110.698—
Greedy GenerationModel=LLaMA 3.1 8B Instruct2025.110.697—
LogTokUModel=LLaMA Guard 32025.110.669—
Marginal ProbabilityModel=LLaMA Guard 22025.110.6580.824
Greedy GenerationModel=LLaMA Guard 22025.110.644—
Conditional ProbabilityModel=LLaMA Guard 22025.110.6440.756
Joint ProbabilityModel=LLaMA Guard 22025.110.6440.754
Joint ProbabilityModel=LLaMA 3.1 8B Instruct2025.110.6260.874
LogTokUModel=LLaMA Guard 22025.110.533—
Probability UncertaintyModel=LLaMA Guard 22025.110.496—
Entropy UncertaintyModel=LLaMA Guard 22025.110.496—
LogTokUModel=LLaMA 3.1 8B Instruct2025.110.462—
Probability UncertaintyModel=LLaMA 3.1 8B Instruct2025.110.453—
Entropy UncertaintyModel=LLaMA 3.1 8B Instruct2025.110.453—