Copyright Tracking on MathV360k subsets of 50k
0.68ASRAGDI
Evaluation Results
| Method | Links | |
|---|---|---|
| AGDIFine-tuning Strategy=LoRA Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.68 | |
| PLAFine-tuning Strategy=LoRA Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.6 | |
| AGDIFine-tuning Strategy=Full Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.45 | |
| OrdinaryFine-tuning Strategy=LoRA Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.41 | |
| PLAFine-tuning Strategy=Full Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.38 | |
| RNAFine-tuning Strategy=LoRA Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.37 | |
| OrdinaryFine-tuning Strategy=Full Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.26 | |
| RNAFine-tuning Strategy=Full Fine-tuning, Base Model=Qwen2-VL [39]2026.02 | 0.21 |