List your company
ResearchFormerly WizWand

General Knowledge QA on PIQA

80AccuracyQwen2.5-7B-Instruct
52.9659.986774.02Jun 12, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct, Ratio=0%, Recovery finetuning=false2026.0680—
Qwen2.5-7B-InstructBackbone=Qwen2.5-7B-Instruct, Ratio=0%, Recovery finetuning=true2026.0680—
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0680—
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0679—
Qwen2.5-3B-InstructBackbone=Qwen2.5-3B-Instruct, Ratio=0%, Recovery finetuning=false2026.0678—
Qwen2.5-3B-InstructBackbone=Qwen2.5-3B-Instruct, Ratio=0%, Recovery finetuning=true2026.0678—
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0678—
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0677—
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0676—
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0676—
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0675—
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0674—
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0674—
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0674—
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0674—
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0674—
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0673—
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0673—
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0671—
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0671—
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0670—
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0670—
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0670—
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0669—
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0669—
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=true2026.0669—
Depth PruningBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0669—
LLM-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0669—
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0668—
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=25%, Recovery finetuning=false2026.0668—
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0668—
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=true2026.0666—
LLM-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.0666—
Adapt-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0666—
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.0665—
Persona-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0665—
Persona-PrunerBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0665—
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=25%, Recovery finetuning=false2026.0664—
Depth PruningBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.0664—
Adapt-PrunerBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.0664—
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=true2026.0664—
SliceGPTBackbone=Qwen2.5-7B-Instruct, Ratio=50%, Recovery finetuning=false2026.0661—
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=true2026.0657—
SliceGPTBackbone=Qwen2.5-3B-Instruct, Ratio=50%, Recovery finetuning=false2026.0654—
Adapt-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—70
Adapt-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—72
Adapt-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—60
Adapt-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—65
Adapt-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—71
Adapt-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—76
Adapt-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—58
Adapt-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—70
Depth PruningRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—70
Depth PruningRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—73
Depth PruningRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—61
Depth PruningRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—66
Depth PruningRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—77
Depth PruningRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—79
Depth PruningRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—67
Depth PruningRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—73
Llama-3.1-8B-InstructRatio=0%, Base Model=Llama-3.1-8B-Instruct2026.06—81
Llama-3.2-3B-InstructRatio=0%, Base Model=Llama-3.2-3B-Instruct2026.06—76
LLM-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—72
LLM-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—73
LLM-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—67
LLM-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—69
LLM-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—78
LLM-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—79
LLM-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—72
LLM-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—74
Persona-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—71
Persona-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—69
Persona-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—64
Persona-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—65
Persona-PrunerRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—75
Persona-PrunerRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—68
Persona-PrunerRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—68
Persona-PrunerRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—65
SliceGPTRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—58
SliceGPTRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—65
SliceGPTRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.2-3B-Instruct2026.06—52
SliceGPTRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.2-3B-Instruct2026.06—55
SliceGPTRatio=25%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—63
SliceGPTRatio=25%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—70
SliceGPTRatio=50%, Recovery finetuning=w/o, Base Model=Llama-3.1-8B-Instruct2026.06—53
SliceGPTRatio=50%, Recovery finetuning=w/, Base Model=Llama-3.1-8B-Instruct2026.06—58