Symbolic Regression on LLM-SRBench OOD (test)
0.325NMSELLM-SR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LLM-SR2026.05 | 0.325 | 42.2 | |
| ICSR2026.05 | 0.87 | 32.7 | |
| ICL2026.05 | 0.883 | 38.3 | |
| IGSR-Agent2026.05 | 1.15 | 62.6 | |
| LaSR2026.05 | 1.5 | 25.1 | |
| IGSRLLM backbone=GPT-4o-mini, Token budget=fixed2026.05 | 1.77 | 1.21 | |
| IGSR-AgentLLM backbone=GPT-4o-mini, Token budget=fixed2026.05 | 2.45 | 1.49 | |
| IGSR2026.05 | 2.9 | 74.4 | |
| LLM-SRLLM backbone=GPT-4o-mini, Token budget=fixed2026.05 | 3.92 | 2.33 | |
| ICSRLLM backbone=GPT-4o-mini, Token budget=fixed2026.05 | 4.22 | 2.71 | |
| ICLLLM backbone=GPT-4o-mini, Token budget=fixed2026.05 | 4.6 | 2.41 | |
| LaSRLLM backbone=GPT-4o-mini, Token budget=fixed2026.05 | 4.77 | 3.33 | |
| D3-w.b.LLM backbone=GPT-4o-mini, Token budget=fixed2026.05 | 6.27 | 4.34 | |
| D3-white-box2026.05 | 29.2 | 4.96 |