Sequential Experimental Design on ChemLex (matched replay)
92.1Final Best PerformanceCARE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CARE2026.06 | 92.1 | 81.6 | 7.9 | |
| Gryffin-style categorical BO2026.06 | 89.8 | 77.2 | 10.2 | |
| EDBO-style descriptor GP-EI2026.06 | 88.9 | 80.5 | 11.1 | |
| Public incumbent2026.06 | 83.9 | 77.2 | 16.1 | |
| Non-evolving LLM policy2026.06 | 81.4 | 76.1 | 18.6 | |
| Ungated LLM-evolving policy2026.06 | 81.2 | 71.2 | 18.8 | |
| LMABO-style LLM-BO2026.06 | 79 | 73.3 | 21 |