DDI Prediction on French Ref.
98.6SensitivityQwen2.5
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen2.5Model Backbone=Qwen-2.5, Model Parameters=3B, Adaptation=Fine-tuned2025.02 | 98.6 | |
| GPT-4oModel Backbone=GPT-4o, Model Parameters=+250B, Adaptation=Fine-tuned2025.02 | 98.1 | |
| Phi3.5Model Backbone=Phi-3.5, Model Parameters=3b, Adaptation=Fine-tuned2025.02 | 97.2 | |
| L2 (Repl.)Model Backbone=L2-regularized logistic regression, Implementation Source=Replication2025.02 | 92.8 | |
| L2 (Paper)Model Backbone=L2-regularized logistic regression, Implementation Source=Original Paper2025.02 | 88.3 | |
| Deepseek QwenModel Backbone=Deepseek-Qwen, Model Parameters=1.5B, Adaptation=Fine-tuned2025.02 | 87.9 | |
| Gemma2Model Backbone=Gemma-2, Model Parameters=9B, Adaptation=Fine-tuned2025.02 | 80.8 |