List your company
ResearchFormerly WizWand

Role-Playing on Alpaca-P

86.66LLM-as-a-Judge ScoreLlama-3.1-8B-Instruct
-3.435219.954943.34566.7351Jun 12, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Llama-3.1-8B-InstructRatio=0%, Base Model=Llama-3.1-8B-Instruct2026.0686.66
Llama-3.2-3B-InstructRatio=0%, Base Model=Llama-3.2-3B-Instruct2026.0684.86
Persona-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0683.9
Persona-PrunerBackbone=Qwen2.5-32B-Instruct, Sparsity=50%2026.0683.1
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0682.61
Persona-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0682.52
Persona-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0681.86
Persona-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0681.73
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0681.34
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0681.1
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0680.45
Persona-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.0680.19
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0678.54
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct, Ratio=0%, Recovery finetuning=false2026.0678.47
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct, Ratio=0%, Recovery finetuning=true2026.0678.47
Persona-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0678.24
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0675.6
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0674.91
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0673.97
Qwen2.5-3B-InstructBackbone=Qwen2.5-3B-Instruct, Ratio=0%, Recovery finetuning=false2026.0671.71
Qwen2.5-3B-InstructBackbone=Qwen2.5-3B-Instruct, Ratio=0%, Recovery finetuning=true2026.0671.71
LLM-PrunerBackbone=Qwen2.5-32B-Instruct, Sparsity=50%2026.0671.01
LLM-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.0670.64
LLM-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0668.59
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0667.79
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.0665.8
Persona-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0664.23
LLM-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0662.99
LLM-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0662.45
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0657.25
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0657
Persona-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.0654.69
Depth PruningRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0653.69
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0653.05
Adapt-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0652.86
Adapt-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0652.63
Adapt-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0651.3
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0651.09
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0650.9
Adapt-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.0650.74
LLM-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0650.37
Depth PruningRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0649.07
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0647.99
Adapt-PrunerBackbone=Qwen2.5-32B-Instruct, Sparsity=50%2026.0647.78
Depth PruningRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0647.3
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0645.17
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0643.52
LLM-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0642.92
LLM-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0642.92
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0642.7
Depth PruningRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0642.04
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0641.74
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0639.64
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0639.07
Adapt-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0636.26
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0636.11
Depth PruningRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0635.6
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0635.37
Adapt-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0634.25
Adapt-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0630.25
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0630.09
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0627.84
LLM-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.0626.6
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0626.41
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0622.19
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0620.93
Depth PruningRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.0620.61
Depth PruningBackbone=Qwen2.5-32B-Instruct, Sparsity=50%2026.0620.12
SliceGPTRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.0619.1
SliceGPTRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.0618.79
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0617.59
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.0615.27
SliceGPTRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.0614.62
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0613.65
Adapt-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.0610.09
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.067.71
SliceGPTRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.067.03
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.066.53
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.065.68
SliceGPTRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.065.32
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.064.66
SliceGPTRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.063.55
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.063.5
Depth PruningRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.062.39
Depth PruningRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.061.3
SliceGPTBackbone=Qwen2.5-32B-Instruct, Sparsity=50%2026.061.04
SliceGPTRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.061.02
SliceGPTRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.060.68
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.060.34
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.060.05
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.060.03