List your company
ResearchFormerly WizWand

Commonsense Reasoning on Commonsense Reasoning Benchmark Intra-domain Multi-task

87.27Average AccuracyST Baseline
83.494884.474985.45586.4351Sep 29, 2025
Updated 6mo ago

Evaluation results

MethodLinks
ST BaselineBackbone=Qwen2-7B2025.0987.27—
ALoRABackbone=Qwen2-7B2025.0986.470.91
LoHaBackbone=Qwen2-7B2025.0986.341.06
MoSLoRABackbone=Qwen2-7B2025.0986.31.09
AdaLoRABackbone=Qwen2-7B2025.0986.141.3
HydraLoRABackbone=Qwen2-7B2025.0986.091.32
LoRABackbone=Qwen2-7B2025.0985.961.43
ST BaselineBackbone=LLaMA3-8B2025.0984.9—
ALoRABackbone=LLaMA3-8B2025.0984.810.04
HydraLoRABackbone=LLaMA3-8B2025.0984.570.32
MoSLoRABackbone=LLaMA3-8B2025.0984.230.76
AdaLoRABackbone=LLaMA3-8B2025.0983.881.17
LoHaBackbone=LLaMA3-8B2025.0983.731.36
LoRABackbone=LLaMA3-8B2025.0983.641.48