List your company

Mathematical Reasoning on Math lm-evaluation-harness (GSM8k, MBPP, MATH)

1.44GSM8k AccuracyTransformer
1.20081.26291.3251.3871Jan 25, 2026

Evaluation results

MethodLinks
TransformerTotal parameters=918m, Non-embedding parameters=764m2026.011.4400.08
Transformer + Spelling Bee EmbeddingsTotal parameters=918m, Non-embedding parameters=764m2026.011.210.20.12