Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Common-sense Reasoning on PIQA, HellaSwag, ARC-C, ARC-E, and OBQA (test)

84.1PIQA Accuracy

LLaMA + LoRA

77.02878.86480.782.536May 23, 2023
Updated 4d ago

Evaluation Results

MethodLinks
2023.05
84.183.359.579.250.671.4
2023.05
82.780.256.875.547.668.6
2023.05
82.582.256.274.94768.6
2023.05
81.880.348.261.642.862.9
2023.05
81.178.853.572.44766.6
2023.05
80.77650.971.64865.5
2023.05
80.378.449.863.342.862.9
2023.05
80.179.245.558.94261.1
2023.05
8078.450.470.847.265.4
2023.05
79.975.246.466.547.263
2023.05
79.676.746.36243.261.5
2023.05
79.475.345.665.84462
2023.05
79.176.244.559.942.260.4
2023.05
78.97446.462.542.860.9
2023.05
78.973.245.165.44461.3
2023.05
78.673.343.755.84358.9
2023.05
77.971.442.457.24258.2
2023.05
77.37341.452.542.457.3