End-to-end interaction performance on GRACE-DS (hidden-test)
75.4E2E Qflexible_iterative
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| flexible_iterativeN=4802026.06 | 75.4 | 77.9 | 60 | 50.1 | 13.6 | 27.5 | 96.9 | |
| flexible_compactN=4802026.06 | 73.4 | 76.1 | 60.7 | 49.2 | 13.7 | 27.5 | 96.5 | |
| restarts_call_matchedN=4802026.06 | 68.6 | 71.6 | 53.1 | 49.3 | 10.5 | 27.8 | 95.8 | |
| restarts_from_scratchN=4802026.06 | 67.2 | 69.7 | 52.1 | 48.5 | 10.5 | 27.8 | 96.5 | |
| fixed_stage_iterativeN=4802026.06 | 65.5 | 77.9 | 60.5 | 48.8 | 14.1 | 27.4 | 84.2 | |
| single_shotN=4802026.06 | 53.6 | 60.4 | 47.5 | 41.4 | 10.5 | 27.6 | 88.8 | |
| unstructured_agentN=4802026.06 | 52.7 | 76.2 | 64.7 | 51.4 | 10.5 | 28.2 | 69.2 |