List your company

Mathematical Reasoning on GSM8k (Acc., log C)

88.7AccuracyBoBN
18.18836.49454.873.106Jul 3, 2026

Evaluation results

MethodLinks
BoBNN=16, Backbone=tulu3-8b2026.0788.7—
BoBNN=4, Backbone=tulu3-8b2026.0786.5—
BoBNN=1, Backbone=tulu3-8b2026.0785.40.43
pi_refN=16, Backbone=tulu3-8b2026.0778.3—
URIALN=16, Backbone=tulu3-8b2026.0773.3—
URIALN=1, Backbone=tulu3-8b2026.0770.60.46
pi_refN=1, Backbone=tulu3-8b2026.0769.60.47
BoBNN=16, Backbone=zephyr-7b2026.0758.7—
BoBNN=4, Backbone=zephyr-7b2026.0744.5—
pi_refN=16, Backbone=zephyr-7b2026.0743.9—
URIALN=16, Backbone=zephyr-7b2026.0741.6—
BoBNN=1, Backbone=zephyr-7b2026.0731.50.63
pi_refN=1, Backbone=zephyr-7b2026.0721.40.72
URIALN=1, Backbone=zephyr-7b2026.0720.90.73