Predicate Grounding on PyBullet Hanoi
100F1 ScoreSYMBOLIZER
Evaluation Results
| Method | Links | |
|---|---|---|
| SYMBOLIZERBackbone=Gemini 3.1 Pro2026.04 | 100 | |
| SYMBOLIZERBackbone=Gemini 3.1 Flash Lite2026.04 | 99.8 | |
| SYMBOLIZERBackbone=Mistral Small2026.04 | 99.4 | |
| SymbolizerModel=Gem.3.1-Pro2026.04 | 95 | |
| SymbolizerModel=Mistral Small2026.04 | 84.1 | |
| SymbolizerModel=Gem.3.1-FL2026.04 | 76.3 | |
| ViLaInBackbone=Gemini 3.1 Flash Lite, Retry Strategy=two retries2026.04 | 64.3 | |
| ViLaInModel=Gem. 3.1-FL, Retries=two retries2026.04 | 29.6 | |
| ViLaInModel=Gem. 3.1-FL, Retries=no retry2026.04 | 13 | |
| ViLaInBackbone=Gemini 3.1 Flash Lite, Retry Strategy=no retry2026.04 | 0 |