Symbolic Regression on NewtonBench Easy
79.2Symbolic AccuracyLLM-AutoSciLab
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LLM-AutoSciLabED (Experiment Design)=true2026.05 | 79.2 | 93.1 | 0.018 | |
| Bayesian OptimizationED (Experiment Design)=true2026.05 | 40.28 | 84.72 | 0.07 | |
| PySRED (Experiment Design)=false2026.05 | 38.89 | 88.89 | 0.03 | |
| Bayesian Experimental DesignED (Experiment Design)=true2026.05 | 19.44 | 78.57 | 0.307 | |
| LLM-onlyED (Experiment Design)=true2026.05 | 16.67 | 19.44 | 3.572 | |
| Code-assisted LLMED (Experiment Design)=true2026.05 | 13.89 | 19.44 | 4.372 |