Law Articles Prediction on JurisMM Text (test)
87.2AccuracyJurisMMA
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| JurisMMABackbone=GPT-4o, Knowledge Base (KB)=true, Multi-agent collaboration (MA)=true2026.01 | 87.2 | 60.4 | 64.5 | 60.2 | |
| TOPJUDGE2026.01 | 76.8 | 45.5 | 41.1 | 40 | |
| JurisMMABackbone=GPT-4o, Knowledge Base (KB)=true, Multi-agent collaboration (MA)=false2026.01 | 76.7 | 59.2 | 60.7 | 56.9 | |
| MPBFN2026.01 | 75.5 | 46.7 | 39.1 | 39.1 | |
| TextCNN2026.01 | 70.3 | 38.9 | 34.3 | 33.4 | |
| JurisMMABackbone=GPT-4o, Knowledge Base (KB)=false, Multi-agent collaboration (MA)=true2026.01 | 65.3 | 32.3 | 32.3 | 28.1 | |
| JurisMMABackbone=Qwen2.5-VL-7B, Knowledge Base (KB)=true, Multi-agent collaboration (MA)=true2026.01 | 62.4 | 26.8 | 31.3 | 26.4 | |
| JurisMMABackbone=Qwen2.5-VL-7B, Knowledge Base (KB)=true, Multi-agent collaboration (MA)=false2026.01 | 57.9 | 25.4 | 29.3 | 25.3 | |
| Qwen2.5-7B-Instruct2026.01 | 52.3 | 13.3 | 17 | 12.3 | |
| JurisMMABackbone=Qwen2.5-VL-7B, Knowledge Base (KB)=false, Multi-agent collaboration (MA)=true2026.01 | 38.9 | 15.9 | 14.3 | 13.2 | |
| GPT-4o-202503262026.01 | 38 | 26.4 | 20.7 | 18.6 | |
| Qwen2.5-VL-7BSFT=true2026.01 | 36.5 | 20.7 | 24.1 | 20 | |
| Qwen2.5-VL-7B-Instruct2026.01 | 27.1 | 13.3 | 12.1 | 9.2 | |
| Qwen2.5-VL-3B-Instruct2026.01 | 22 | 8.9 | 7.2 | 5.7 | |
| GLM-4V-9B2026.01 | 9.5 | 1.9 | 2.7 | 1.5 | |
| mPLUG-7B2026.01 | 9.4 | 2.7 | 3.5 | 1.9 |