Cyberattack Detection on TCDR (test)
97.62Detection RateDistilBERT
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DistilBERTModel Type=LLMs, Fine-tuned=true, Base Model=distilbert-base-uncased2026.01 | 97.62 | 99.84 | 100 | 98.81 | 100 | 99.36 | |
| GPT-2Model Type=LLMs, Fine-tuned=true2026.01 | 97.06 | 99.8 | 99.9 | 98.53 | 100 | 99.2 | |
| GRUModel Type=Non-LLM Models2026.01 | 96.98 | 99.78 | 99.88 | 98.49 | 100 | 99.17 | |
| CNNModel Type=Non-LLM Models2026.01 | 96.48 | 99.74 | 99.86 | 98.24 | 100 | 99.03 | |
| Random ForestModel Type=Non-LLM Models2026.01 | 95.73 | 99.68 | 99.83 | 97.86 | 100 | 98.82 | |
| KNNModel Type=Non-LLM Models2026.01 | 95.6 | 99.66 | 99.76 | 97.8 | 100 | 98.75 | |
| XGBoostModel Type=Non-LLM Models2026.01 | 94.85 | 99.62 | 99.79 | 97.42 | 100 | 98.57 | |
| LSTMModel Type=Non-LLM Models2026.01 | 94.35 | 99.58 | 99.77 | 97.17 | 100 | 98.43 | |
| Decision TreeModel Type=Non-LLM Models2026.01 | 93.59 | 98.56 | 96.89 | 96.71 | 99.8 | 97.61 | |
| SVMModel Type=Non-LLM Models2026.01 | 92.34 | 99.43 | 99.69 | 96.17 | 100 | 97.85 | |
| DistilBERT+LoRAModel Type=LLMs, Tuning Method=Low-Rank Adaptation (LoRA), Rank (r)=8, Alpha=322026.01 | 92.31 | 99.49 | 99.73 | 96.15 | 100 | 97.86 | |
| Naive BayesModel Type=Non-LLM Models2026.01 | 69.85 | 85.05 | 63.14 | 78.06 | 78.06 | 66.22 | |
| Logistic RegressionModel Type=Non-LLM Models2026.01 | 68.22 | 97.64 | 98.75 | 84.11 | 100 | 89.92 |