Formal Language Generation on CPP-Bench
97Syntactic Accuracy@1LAVE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| LAVEModel backbone=DiffuC-7B2026.01 | 97 | 99.1 | 99.6 | 100 | |
| LAVEModel backbone=Dream-7B2026.01 | 96 | 98.1 | 99.4 | 99.4 | |
| FS-CDModel backbone=DiffuC-7B2026.01 | 93.8 | 94 | 94 | 94.3 | |
| IG-CDModel backbone=DiffuC-7B2026.01 | 93.8 | 94.5 | 95.1 | 95.3 | |
| LAVEModel backbone=LLaDA-1.52026.01 | 91.3 | 97.6 | 98.8 | 98.8 | |
| LAVEModel backbone=LLaDA-8B2026.01 | 90.9 | 98.7 | 99.3 | 99.4 | |
| IG-CDModel backbone=LLaDA-1.52026.01 | 89.5 | 96.1 | 97.3 | 98.8 | |
| IG-CDModel backbone=Dream-7B2026.01 | 87.6 | 90.2 | 90.4 | 90.9 | |
| IG-CDModel backbone=LLaDA-8B2026.01 | 86.1 | 92.6 | 96.7 | 98.1 | |
| FS-CDModel backbone=Dream-7B2026.01 | 83.1 | 84.3 | 84.4 | 84.4 | |
| NO-CDModel backbone=DiffuC-7B2026.01 | 79.9 | 80.2 | 80.5 | 81.7 | |
| NO-CDModel backbone=LLaDA-1.52026.01 | 78.2 | 89.6 | 92 | 94.8 | |
| NO-CDModel backbone=Dream-7B2026.01 | 76.2 | 76.8 | 79.2 | 81.1 | |
| NO-CDModel backbone=LLaDA-8B2026.01 | 74.2 | 90.5 | 94.1 | 95.6 | |
| FS-CDModel backbone=LLaDA-1.52026.01 | 73.8 | 74 | 74.2 | 74.2 | |
| FS-CDModel backbone=LLaDA-8B2026.01 | 66.3 | 66.7 | 66.7 | 66.7 |