Tool Shortlisting on AppWorld v1.0 (test)
0.71R-precision (AZ)TabAgent
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TabAgentTraining Supervision=Real+synth2026.02 | 0.71 | 0.66 | 0.9 | 0.96 | 0.61 | 0.95 | 0.93 | 0.98 | 1 | 0.88 | 0.97 | 0.95 | 1 | 1 | 0.92 | |
| DSR-FTEvaluation Protocol=Fine-tuned, Training Supervision=Real+synth2026.02 | 0.69 | 0.56 | 0.85 | 0.92 | 0.64 | 0.93 | 0.9 | 0.98 | 1 | 0.85 | 0.95 | 0.93 | 0.99 | 1 | 0.88 | |
| TabAgentTraining Supervision=Real-Only2026.02 | 0.57 | 0.58 | 0.66 | 0.62 | 0.7 | 0.91 | 0.88 | 0.98 | 1 | 0.88 | 0.93 | 0.91 | 1 | 1 | 0.9 | |
| Llama-3.1-8B-InstructBackbone=Llama-3.1-8B, Evaluation Protocol=Zero-Shot2026.02 | 0.56 | 0.45 | 0.38 | 0.7 | 0.67 | 0.56 | 0.5 | 0.65 | 0.71 | 0.67 | 0.56 | 0.5 | 0.65 | 0.73 | 0.69 | |
| DSR-FTEvaluation Protocol=Fine-tuned, Training Supervision=Real-Only2026.02 | 0.45 | 0.52 | 0.57 | 0.63 | 0.46 | 0.89 | 0.84 | 0.95 | 0.99 | 0.82 | 0.9 | 0.87 | 0.95 | 0.99 | 0.85 | |
| DSREvaluation Protocol=Zero-Shot2026.02 | 0.39 | 0.18 | 0.3 | 0.55 | 0.3 | 0.65 | 0.46 | 0.82 | 0.89 | 0.62 | 0.67 | 0.53 | 0.88 | 0.96 | 0.7 | |
| Llama-3.2-3B-InstructBackbone=Llama-3.2-3B, Evaluation Protocol=Zero-Shot2026.02 | 0.38 | 0.34 | 0.4 | 0.42 | 0.39 | 0.6 | 0.58 | 0.63 | 0.76 | 0.64 | 0.66 | 0.66 | 0.66 | 0.79 | 0.67 | |
| BM25Evaluation Protocol=Zero-Shot2026.02 | 0.28 | 0.19 | 0.35 | 0.42 | 0.28 | 0.48 | 0.51 | 0.54 | 0.72 | 0.6 | 0.54 | 0.58 | 0.57 | 0.74 | 0.71 | |
| Llama-3.2-1B-InstructBackbone=Llama-3.2-1B, Evaluation Protocol=Zero-Shot2026.02 | 0.02 | 0.03 | 0.06 | 0.07 | 0.09 | 0.29 | 0.27 | 0.29 | 0.45 | 0.33 | 0.32 | 0.29 | 0.33 | 0.46 | 0.34 |