In-distribution
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Harmfulness Detection | Full in-distribution (test) | AUROC0.964 | 63 | |
| Error Detection | In-distribution (test) | AUC0.8916 | 40 | |
| Mathematical Reasoning | In-Distribution Avg | Average Score45.6 | 29 | |
| Visual Grounding | In-Distribution (test) | Accuracy99.7 | 18 | |
| Debiasing Effectiveness | In-Distribution (ID) | Mean Effectiveness Score (ID)10.2 | 16 | |
| Agentic Model Routing | In-Distribution n=2,919 (test) | Average Performance (%)57 | 15 | |
| LLM Negotiation | In-distribution (held-out) | Reward0.58 | 13 | |
| Trajectory Prediction | In-distribution (InDist) (test) | ADE0.004 | 10 | |
| Object type prediction | In-Distribution (ID) | Accuracy (ID)100 | 9 | |
| Reasoning step reduction | In-Distribution 5K corpus (test) | Savings Rate47.5 | 9 | |
| Policy Evaluation | In-distribution | r Score98.4 | 8 | |
| Interactive Segmentation | In-distribution | NoC@90 Clicks1.95 | 6 | |
| Point Tracking | In-distribution | Avg Displacement Error57.4 | 6 | |
| Edge-selection fidelity | in-distribution clean (test) | F1 Score53.4 | 5 | |
| Prompt Injection Detection | In-distribution (ID) (test) | Macro F1 Score95.41 | 5 | |
| Text-to-Speech | In-distribution ID (test) | MOS3.87 | 5 | |
| Metasurface inverse design | In-Distribution (test) | SG74 | 2 |
Showing 17 of 17 rows