Tool Sequence Prediction on HuggingFace
89.99Tool F1GRAFT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GRAFTLLM=Mistral-7B2026.05 | 89.99 | 13 | 85 | 65 | |
| GRAFTLLM=Llama-3.2-3B2026.05 | 89.93 | 13 | 84 | 65 | |
| ToolGenLLM=Mistral-7B2026.05 | 88.04 | 14 | 82 | 62 | |
| ToolGenLLM=Llama-3.2-3B2026.05 | 86.75 | 17 | 81 | 57 | |
| GRAFTLLM=Qwen2.5-1.5B2026.05 | 86.5 | 18 | 79 | 57 | |
| ToolGenLLM=Qwen2.5-1.5B2026.05 | 85.98 | 19 | 78 | 54 | |
| GToolLLM=Mistral-7B2026.05 | 82.85 | 23 | 72 | 49 | |
| GToolLLM=Llama-3.2-3B2026.05 | 77.81 | 30 | 69 | 41 | |
| GToolLLM=Qwen2.5-1.5B2026.05 | 67.21 | 43 | 64 | 25 | |
| GNN4PlanLLM=Llama-3.2-3B2026.05 | 63.66 | 47 | 53 | 23 | |
| BeamSearchLLM=Mistral-7B2026.05 | 61.78 | 49 | 46 | 14 | |
| GNN4PlanLLM=Qwen2.5-1.5B2026.05 | 58.82 | 52 | 47 | 16 | |
| GNN4PlanLLM=Mistral-7B2026.05 | 58.78 | 51 | 51 | 18 | |
| PLaGLLM=Mistral-7B2026.05 | 57.08 | 54 | 48 | 9 | |
| DirectLLM=Mistral-7B2026.05 | 55.39 | 56 | 45 | 8 | |
| HuggingGPTLLM=Mistral-7B2026.05 | 55.29 | 56 | 46 | 8 | |
| BeamSearchLLM=Llama-3.2-3B2026.05 | 49.18 | 62 | 38 | 8 | |
| HuggingGPTLLM=Llama-3.2-3B2026.05 | 48.55 | 63 | 36 | 10 | |
| DirectLLM=Llama-3.2-3B2026.05 | 48.5 | 61 | 39 | 10 | |
| PLaGLLM=Llama-3.2-3B2026.05 | 48.18 | 71 | 35 | 8 | |
| DirectLLM=Qwen2.5-1.5B2026.05 | 38.21 | 70 | 28 | 5 | |
| BeamSearchLLM=Qwen2.5-1.5B2026.05 | 35.59 | 76 | 28 | 2 | |
| HuggingGPTLLM=Qwen2.5-1.5B2026.05 | 32.68 | 74 | 20 | 3 | |
| PLaGLLM=Qwen2.5-1.5B2026.05 | 29.92 | 76 | 21 | 5 |