SFT Generalization on Tulu3 SFT (Original)
79.9General ScoreQwen-7B + Tulu-Delta
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen-7B + Tulu-DeltaBase Model=Qwen2.5-7B, Alignment Delta=Tulu32026.05 | 79.9 | |
| Dense-FT-AvgBase Model=Qwen2.5-7B, Alignment Delta=Tulu3, Comparison Category=Baselines w/ Same CPT Data2026.05 | 76.46 | |
| DeltaMoEBase Model=Qwen2.5-7B, Alignment Delta=Tulu3, Comparison Category=Baselines w/ Matched Parameters2026.05 | 76.14 | |
| LLaMA-ProBase Model=Qwen2.5-7B, Alignment Delta=Tulu3, Comparison Category=Baselines w/ Matched Parameters2026.05 | 75.49 | |
| Dense-FT-Avg-2FLOPsBase Model=Qwen2.5-7B, Alignment Delta=Tulu3, Comparison Category=Baselines w/ Matched FLOPs (2x CPT Data)2026.05 | 75.27 | |
| MoLABase Model=Qwen2.5-7B, Alignment Delta=Tulu3, Comparison Category=Baselines w/ Matched Parameters2026.05 | 74.59 | |
| Dense-FT-DeltaBase Model=Qwen2.5-7B, Alignment Delta=Tulu3, Comparison Category=Baselines w/ Same CPT Data2026.05 | 72.6 | |
| Dense-FT-Delta-2FLOPsBase Model=Qwen2.5-7B, Alignment Delta=Tulu3, Comparison Category=Baselines w/ Matched FLOPs (2x CPT Data)2026.05 | 68.92 |