Predicate Grounding on ProDG Hanoi
97.3F1 ScoreSymbolizer
Evaluation Results
| Method | Links | |
|---|---|---|
| SymbolizerModel=Gem.3.1-FL2026.04 | 97.3 | |
| SymbolizerModel=Gem.3.1-Pro2026.04 | 88.6 | |
| SymbolizerModel=Mistral Small2026.04 | 78.4 | |
| ViLaInModel=Gem. 3.1-FL, Retries=no retry2026.04 | 65.8 | |
| ViLaInModel=Gem. 3.1-FL, Retries=two retries2026.04 | 64.1 | |
| ViLaInModel=GPT-4, Retries=two retries2026.04 | 52 | |
| ViLaInModel=GPT-4, Retries=no retry2026.04 | 47 | |
| SYMBOLIZERBackbone=Gemini 3.1 Flash Lite2026.04 | 0.999 | |
| SymbolizerModel=Gem.3.1-Pro2026.04 | 0.986 | |
| SYMBOLIZERBackbone=Gemini 3.1 Pro2026.04 | 0.943 | |
| SYMBOLIZERBackbone=Mistral Small2026.04 | 0.927 | |
| ViLaInBackbone=Gemini 3.1 Flash Lite, Retry Strategy=two retries2026.04 | 0.9 | |
| SymbolizerModel=Gem.3.1-FL2026.04 | 0.898 | |
| ViLaInBackbone=Gemini 3.1 Flash Lite, Retry Strategy=no retry2026.04 | 0.897 | |
| ViLaInBackbone=GPT-4, Retry Strategy=no retry2026.04 | 0.89 | |
| ViLaInBackbone=GPT-4, Retry Strategy=two retries2026.04 | 0.89 | |
| SymbolizerModel=Mistral Small2026.04 | 0.846 | |
| ViLaInModel=Gem. 3.1-FL, Retries=two retries2026.04 | 0.279 | |
| ViLaInModel=Gem. 3.1-FL, Retries=no retry2026.04 | 0.011 |