List your company

Grade School Math Reasoning on GSM8K No CoT augmented (test)

0.2482AccuracyTurboConn
0.0649520.1125260.16010.207674Feb 20, 2026

Evaluation results

MethodLinks
TurboConnModel=Llama 3.1 8B2026.020.2482
BaselineModel=Llama 3.1 8B2026.020.2392
TurboConnModel=Qwen 3 1.7B2026.020.2031
BaselineModel=Qwen 3 1.7B2026.020.159
TurboConnModel=Llama 3.2 1B2026.020.0832
BaselineModel=Llama 3.2 1B2026.020.072