Multi-domain Science Q&A on SciKnowEval Mat-Phys (test)
70.5Accuracy (Mat. Sci.)OPCoD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OPCoDBase Student=Student 1, Domain Strength=Mat-stronger2026.06 | 70.5 | 55.3 | 62.9 | |
| GRPOBase Student=Student 1, Domain Strength=Mat-stronger2026.06 | 67.8 | 53.8 | 60.8 | |
| OPCoDBase Student=Student 2, Domain Strength=Phys-stronger2026.06 | 66.1 | 54.3 | 60.2 | |
| GRPOBase Student=Student 2, Domain Strength=Phys-stronger2026.06 | 66 | 53.9 | 60 | |
| Student 1Domain Strength=Mat-stronger2026.06 | 65.1 | 48.1 | 56.6 | |
| SDPOBase Student=Student 1, Domain Strength=Mat-stronger2026.06 | 62.2 | 52.3 | 57.2 | |
| SDPOBase Student=Student 2, Domain Strength=Phys-stronger2026.06 | 58.8 | 51.9 | 55.4 | |
| Student 2Domain Strength=Phys-stronger2026.06 | 56 | 51.6 | 53.8 |