Science Question Answering on Chemistry
80.6Mean@16SC-SDPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SC-SDPOBackbone=Qwen3-8B, Alpha=0.52026.05 | 80.6 | 81.3 | |
| SRPOBackbone=Qwen3-8B2026.05 | 80.5 | 81.6 | |
| SDPOBackbone=Qwen3-8B, Filtering Strategy=Hard Filter2026.05 | 80 | 80.4 | |
| SC-SDPOBackbone=Qwen3-8B, Alpha=12026.05 | 80 | 79.5 | |
| SDPOBackbone=Qwen3-8B2026.05 | 78.7 | 79.2 | |
| SDPOBackbone=Qwen3-8B, Weighting Strategy=PACED2026.05 | 74.7 | 75.5 | |
| GRPOBackbone=Qwen3-8B2026.05 | 68.5 | 70.9 | |
| GRPOBackbone=Qwen3-8B, Normalization=None2026.05 | 63.9 | 67 | |
| Qwen3-8BBackbone=Qwen3-8B2026.05 | 42.4 | 35.7 |