Malicious Code Detection on NPM (test)
99.66PrecisionCerebro
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Cerebro2026.01 | 99.66 | 87.32 | 72.11 | 83.68 | 2 | 226 | |
| Cerebro RoBERTaTrain=Mixed2023.09 | 98.9 | — | 93.9 | — | — | — | |
| Cerebro T5Train=Mixed2023.09 | 98.9 | — | 92 | — | — | — | |
| Cerebro T5Train=NPM, Backbone=T5, Learning-based=true2023.09 | 98.9 | — | 91 | — | — | — | |
| Cerebro BERTTrain=Mixed2023.09 | 98.7 | — | 93 | — | — | — | |
| Cerebro RoBERTaTrain=NPM, Backbone=RoBERTa, Learning-based=true2023.09 | 98.5 | — | 92.9 | — | — | — | |
| PYGUARDTechnique=RAG, Backbone=GPT-4.12026.01 | 98.26 | 98.07 | 97.53 | 97.89 | 14 | 20 | |
| Cerebro BERTTrain=NPM, Backbone=BERT, Learning-based=true2023.09 | 98.2 | — | 91.8 | — | — | — | |
| GPT-4.12026.01 | 98.08 | 96.71 | 94.68 | 96.35 | 15 | 43 | |
| GuardDog2026.01 | 97.44 | 94.1 | 89.49 | 93.3 | 19 | 85 | |
| AmalfiNBTrain=Mixed2023.09 | 93.9 | — | 76.6 | — | — | — | |
| AmalfiDTTrain=NPM, Algorithm=Decision Tree, Learning-based=true2023.09 | 92.7 | — | 86 | — | — | — | |
| AmalfiDTTrain=Mixed2023.09 | 92.1 | — | 85.8 | — | — | — | |
| AmalfiNBTrain=NPM, Algorithm=Naive Bayes, Learning-based=true2023.09 | 92.1 | — | 77.1 | — | — | — | |
| SAPTrain=NPM, Learning-based=true2023.09 | 90.5 | — | 35.4 | — | — | — | |
| SAPTrain=Mixed2023.09 | 90.4 | — | 39.5 | — | — | — | |
| OSSGadget2026.01 | 49.83 | 53.8 | 90.11 | 64.17 | 734 | 80 | |
| AmalfiSVMTrain=Mixed2023.09 | 28.8 | — | 94.6 | — | — | — | |
| AmalfiSVMTrain=NPM, Algorithm=SVM, Learning-based=true2023.09 | 10.2 | — | 26.7 | — | — | — |