Multi-stage S+P Tcl generation on AutoEDA Tcl Benchmark 10k prompt-Tcl configuration pairs
15,292.71Token CountBaseline 2
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Baseline 2Generation Strategy=In-Context Learning, Base Model=gpt-4o2025.08 | 15,292.71 | 98.634 | 98.993 | |
| Baseline 1Generation Strategy=Direct Generation, Base Model=gpt-4o2025.08 | 2,600.78 | 0.218 | 6.328 | |
| AutoEDAGeneration Strategy=Microservice-Based LLM Agents, Base Model=Qwen3-0.6B2025.08 | 460.66 | 98.633 | 99.023 |