Driving Quality Classification on DriveJudge (val)
80.97Accuracy (tau=0.7)DriveJudge-2B
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DriveJudge-2BFinetuning=Yes2026.06 | 80.97 | 83.28 | 83.74 | 77.95 | 58.31 | 89.83 | |
| DriveJudge-8BFinetuning=Yes2026.06 | 80.89 | 83.47 | 84.03 | 78.9 | 59.04 | 90.77 | |
| Supervised ClassifierFinetuning=Yes2026.06 | 77.01 | 73.59 | 33.51 | 58.39 | 33.28 | 80.49 | |
| GPT-5.4Finetuning=No2026.06 | 76.71 | 76.59 | 72.48 | 48.59 | 23.65 | 76.89 | |
| EPDMSFinetuning=No2026.06 | 66.94 | 54.76 | 43.95 | 57.67 | 26.34 | 81.64 | |
| Gemini3.1Finetuning=No2026.06 | 65.73 | 65.69 | 65.26 | 56.61 | 26.29 | 79.62 | |
| RandomFinetuning=No2026.06 | 50 | 50 | 50 | 50 | 23.3 | 76.7 |