Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Zero-shot Language Understanding on BoolQ, ARC-e, ARC-c, WinoGrande, and HellaSwag

83.08ARC-e Accuracy

MoEITS

72.700875.395478.0980.7846Apr 12, 2026
Updated 5d ago

Evaluation Results

MethodLinks
2026.04
83.0887.3355.8980.9882.8378.02
2026.04
82.3786.5175.4355.874.2874.88
2026.04
81.98653.375.562.371.8
2026.04
80.0383.73-72.3756.1673.1
2026.04
80-51.5-59.963.8
2026.04
79.984.853.973.860.170.5
2026.04
78.282.652.271.561.169.1
2026.04
73.180.648.471.374.969.6