Mathematical Reasoning on MATH 500 (Accuracy, log C)
54.4AccuracyBoBN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BoBNN=16, Backbone=tulu3-8b2026.07 | 54.4 | — | |
| pi_refN=16, Backbone=tulu3-8b2026.07 | 53.8 | — | |
| URIALN=16, Backbone=tulu3-8b2026.07 | 52 | — | |
| BoBNN=4, Backbone=tulu3-8b2026.07 | 49.6 | 0.55 | |
| BoBNN=1, Backbone=tulu3-8b2026.07 | 46.6 | 0.55 | |
| pi_refN=1, Backbone=tulu3-8b2026.07 | 45.6 | 0.58 | |
| URIALN=1, Backbone=tulu3-8b2026.07 | 42 | 0.55 | |
| BoBNN=16, Backbone=zephyr-7b2026.07 | 21 | — | |
| pi_refN=16, Backbone=zephyr-7b2026.07 | 17.4 | — | |
| BoBNN=4, Backbone=zephyr-7b2026.07 | 15.2 | — | |
| URIALN=16, Backbone=zephyr-7b2026.07 | 15 | — | |
| BoBNN=1, Backbone=zephyr-7b2026.07 | 13.8 | 0.63 | |
| URIALN=1, Backbone=zephyr-7b2026.07 | 12.2 | 0.66 | |
| pi_refN=1, Backbone=zephyr-7b2026.07 | 10.2 | 0.67 |