List your company

Zero-shot Commonsense Reasoning Suite (PIQA, HellaSwag, WinoGrande, ARC-E/C, SIQA, BoolQ, LAMBADA)

73.3PIQA Accuracy (Zero-shot)OSDN-APF
72.67672.8387373.162May 13, 2026

Evaluation results

MethodLinks
OSDN-APFModel Size=1.3B, Training Tokens=100B, zero-shot=true2026.0573.359.260.66940.441.760.54856.6
DeltaNetModel Size=1.3B, Training Tokens=100B, zero-shot=true2026.0572.758.159.868.741.841.557.947.356