Symbolic Regression on NewtonBench Medium
72.2Symbolic AccuracyLLM-AutoSciLab
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LLM-AutoSciLabED (Experiment Design)=true2026.05 | 72.2 | 84.7 | 0.04 | |
| PySRED (Experiment Design)=false2026.05 | 27.78 | 75 | 0.125 | |
| Bayesian OptimizationED (Experiment Design)=true2026.05 | 27.78 | 70.83 | 0.212 | |
| Bayesian Experimental DesignED (Experiment Design)=true2026.05 | 11.11 | 66.18 | 0.449 | |
| Code-assisted LLMED (Experiment Design)=true2026.05 | 5.56 | 11.11 | 5.437 | |
| LLM-onlyED (Experiment Design)=true2026.05 | 2.78 | 2.78 | 4.738 |