Symbolic Regression on LSR-Synth Chemistry
13.89SA (%)PiT-PO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PiT-POBackbone=Llama-3.1-8B-Instruct2026.02 | 13.89 | 77.78 | 0 | |
| uDSR+Optimization Strategy=SAGE-Fit (Inner-loop)2026.05 | 13.89 | 64.89 | 0 | |
| uDSROptimization Strategy=Post-hoc Refinement2026.05 | 11.11 | 54.29 | 0 | |
| LLM-SR+Backbone=Qwen2.5-7B, Optimization Strategy=SAGE-Fit (Inner-loop)2026.05 | 11.11 | 60.61 | 0 | |
| LLM-SRBackbone=Llama-3.1-8B-Instruct2026.02 | 8.33 | 66.66 | 0 | |
| PySR+Optimization Strategy=SAGE-Fit (Inner-loop)2026.05 | 8.33 | 43.33 | 0.0004 | |
| LLM-SRBackbone=Qwen2.5-7B, Optimization Strategy=Post-hoc Refinement2026.05 | 8.33 | 33.33 | 0 | |
| LaSR+Backbone=Qwen2.5-7B, Optimization Strategy=SAGE-Fit (Inner-loop)2026.05 | 5.56 | 27.78 | 0.0014 | |
| PySROptimization Strategy=Post-hoc Refinement2026.05 | 2.78 | 37.14 | 0.0005 | |
| LaSRBackbone=Qwen2.5-7B, Optimization Strategy=Post-hoc Refinement2026.05 | 2.78 | 25 | 0.0015 | |
| Direct PromptingBackbone=Llama-3.1-8B-Instruct2026.02 | 0 | 0 | 0.0644 | |
| SGABackbone=Llama-3.1-8B-Instruct2026.02 | 0 | 8.33 | 0.0458 | |
| LaSRBackbone=Llama-3.1-8B-Instruct2026.02 | 0 | 27.77 | 0.0003 |