Civil Judgment Generation on Chinese Civil Court Simulation dataset (test)
5.46JCCGLM-Sim
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GLM-SimBackbone=GLM, Framework=Civil Court Simulation2026.06 | 5.46 | 5.96 | 4.45 | 5.59 | 5.34 | 5.31 | |
| Kimi-FactBackbone=Kimi, Framework=Direct Generation (Fact)2026.06 | 5.31 | 5.44 | 4.45 | 5.54 | 4.98 | 5.11 | |
| Kimi-SimBackbone=Kimi, Framework=Civil Court Simulation2026.06 | 5.25 | 5.86 | 4.47 | 5.6 | 5.05 | 5.19 | |
| DS-SimBackbone=DeepSeek, Framework=Civil Court Simulation2026.06 | 5.15 | 5.71 | 4.21 | 5.25 | 5.06 | 5.03 | |
| GLM-FactBackbone=GLM, Framework=Direct Generation (Fact)2026.06 | 5.15 | 5.55 | 4.15 | 5.34 | 4.94 | 4.98 | |
| DS-FactBackbone=DeepSeek, Framework=Direct Generation (Fact)2026.06 | 4.86 | 5.26 | 3.77 | 5.12 | 4.78 | 4.68 |