List your company

Commonsense Reasoning on Common

65.69AccuracyKALE
57.411659.560861.7163.8592Jan 12, 2026

Evaluation results

MethodLinks
KALEBackbone=Llama3 8B2026.0165.69
COT-DistillBackbone=Llama3 8B2026.0161.64
Distill-SBSBackbone=Llama3 8B2026.0160.85
PDBackbone=Llama3 8B2026.0157.73