Mathematical Reasoning on GSM8K (Deductive, Inductive, Abductive)
79.37Deductive AccuracyComplementary S.
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Complementary S.Backbone=Llama-3.1-8B-it2026.04 | 79.37 | 78.33 | 78.62 | |
| Mono S.Backbone=Llama-3.1-8B-it2026.04 | 78.31 | 77.75 | 78.24 | |
| UnsteeredBackbone=Llama-3.1-8B-it2026.04 | 76.26 | 76.26 | 76.26 |