Research map / Large language models / Agents, code and math
GTool: Graph Enhanced Tool Planning with Large Language Model
Wenjie Chen, Di Yao, Wenbin Li, Xuying Meng, Chang Gong, Jingping Bi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add request-specific node to GTool (ablation) | GTool · HuggingFace | n-F1: 0.7128 -> 0.7913 | Improved |
| add missing dependency prediction (MDPL) to GTool (ablation) | GTool · HuggingFace | n-F1: 0.7650 -> 0.7913 | Improved |
| add missing dependency prediction (MDPL) to GTool (ablation) | GTool · HuggingFace | NED: 0.2541 -> 0.2537 | No clear effect |
| add both request-specific node and MDPL to GTool (ablation) | GTool · HuggingFace | n-F1: 0.6131 -> 0.7913 | Improved |
| add tool graph token to LLM prompt (ablation) | GTool · HuggingFace | n-F1: 0.1566 -> 0.7913 | Improved |
| add request-specific node on Vicuna-13B (ablation) | GTool · HuggingFace | n-F1: 0.7589 -> 0.8029 | Improved |
| add MDPL on Vicuna-13B (ablation) | GTool · HuggingFace | n-F1: 0.7707 -> 0.8029 | Improved |
| use GTool graph token instead of GNN4Plan | GNN4Plan · HuggingFace | n-F1: 0.4853 -> 0.7913 | Improved |
| use GTool instead of HuggingGPT on ToolE with Vicuna-13B | HuggingGPT · ToolE | l-F1: 0.3740 -> 0.3500 | Hurt |
| use GTool instead of HuggingGPT on ToolBench | HuggingGPT · ToolBench | n-F1: 0.5989 -> 0.6126 | Improved |
| increase MDPL sample size alpha from 4 to 6 (ablation) | GTool · TaskBench | n-F1: 0.13% gain | No clear effect |
| increase loss weight lambda beyond 0.1 (ablation) | GTool · TaskBench | performance: peaks at lambda=0.1 then degrades | Mixed |
Closest papers by meaning
- ToolTree: Efficient LLM Tool Planning via Dual-Feedback Monte Carlo Tree Search and Bidirectional Pruning (ICLR 2026 · similarity 0.90)
- Tool-Planner: Task Planning with Clusters across Multiple Tools (ICLR 2025 · similarity 0.89)
- Large Language Models as Tool Makers (ICLR 2024 · similarity 0.87)
- GTA: A Benchmark for General Tool Agents (NeurIPS 2024 · similarity 0.87)
- Joint Optimization of Tool Creation and Use for Large Language Model Agents (NeurIPS 2026 · similarity 0.86)
- ToolGen: Unified Tool Retrieval and Calling via Generation (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
