LTL Instruction Following on ChessWorld finite-horizon phi4 (test)
0.927SRLTL-GNN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LTL-GNNHorizon=Finite2025.12 | 0.927 | 0.902 | |
| DeepLTLHorizon=Finite2025.12 | 0.912 | 0.886 | |
| TransformerHorizon=Finite2025.12 | 0.77 | 0.746 |
| Method | Links | ||
|---|---|---|---|
| LTL-GNNHorizon=Finite2025.12 | 0.927 | 0.902 | |
| DeepLTLHorizon=Finite2025.12 | 0.912 | 0.886 | |
| TransformerHorizon=Finite2025.12 | 0.77 | 0.746 |