List your company
ResearchFormerly WizWand

Language Modeling on Penn Treebank

11.18PerplexityMeta-Llama-3-8B Dense
7.825230.470153.11575.7599Mar 16, 2024Jul 13, 2024Nov 9, 2024Mar 9, 2025Jul 6, 2025Nov 2, 2025Mar 2, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Meta-Llama-3-8B DenseConfig=Dense, Fine-tuning Protocol=None, Base Model=Meta-Llama-3-8B2026.0311.18
LFT-3BASiL-TMConfig=4:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0313.41
LFT-Hassle-free-ALPS-TMConfig=4:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0313.43
LFT-3BASiLConfig=4:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0313.43
LFT-Hassle-free-SparseGPT-TMConfig=4:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0313.71
LFT-Hassle-free-SparseGPTConfig=4:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.03
LFT-3BASiL-TMConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.17
LFT-Hassle-free-ALPS-TMConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.34
LFT-Hassle-free-ALPSConfig=4:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.45
LFT-3BASiLConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.54
LFT-Hassle-free-ALPSConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.61
LFT-OATSConfig=4:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.64
LFT-OATS-TMConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.81
LFT-Hassle-free-SparseGPT-TMConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0314.84
LFT-Hassle-free-SparseGPTConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0315.35
LFT-OATSConfig=2:4+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0316.24
LFT-3BASiL-TMConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0317.27
Llama-3.2-1B DenseConfig=Dense2026.0317.59
LFT-Hassle-free-ALPS-TMConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0317.74
LFT-3BASiLConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0317.85
LFT-Hassle-free-ALPSConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0317.89
LFT-Hassle-free-SparseGPT-TMConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0318.08
LFT-OATS-TMConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0318.34
LFT-Hassle-free-SparseGPTConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0318.84
LLaMA-13B (Ma, Fang, and Wang 2023)Pruning Ratio=0%, Zero-shot=true2024.0320.24
LFT-OATSConfig=3:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0320.59
LFT-3BASiL-TMConfig=4:8+64LR, Fine-Tuning=LoRA2026.0322.3
LFT-3BASiLConfig=4:8+64LR, Fine-Tuning=LoRA2026.0322.56
LFT-Hassle-free-ALPS-TMConfig=4:8+64LR, Fine-Tuning=LoRA2026.0322.75
LFT-Hassle-free-SparseGPT-TMConfig=4:8+64LR, Fine-Tuning=LoRA2026.0323.6
LFT-Hassle-free-ALPSConfig=4:8+64LR, Fine-Tuning=LoRA2026.0323.63
LFT-OATS-TMConfig=4:8+64LR, Fine-Tuning=LoRA2026.0323.65
LFT-3BASiL-TMConfig=2:4+64LR, Fine-Tuning=LoRA2026.0323.88
LFT-Hassle-free-ALPS-TMConfig=2:4+64LR, Fine-Tuning=LoRA2026.0324.31
LFT-3BASiLConfig=2:4+64LR, Fine-Tuning=LoRA2026.0324.56
LFT-Hassle-free-SparseGPTConfig=4:8+64LR, Fine-Tuning=LoRA2026.0324.74
LFT-Hassle-free-SparseGPT-TMConfig=2:4+64LR, Fine-Tuning=LoRA2026.0325.23
LFT-Hassle-free-ALPSConfig=2:4+64LR, Fine-Tuning=LoRA2026.0325.47
LFT-OATS-TMConfig=2:4+64LR, Fine-Tuning=LoRA2026.0325.77
LFT-Hassle-free-SparseGPTConfig=2:4+64LR, Fine-Tuning=LoRA2026.0326.81
HyWIAPruning Ratio=20%, Zero-shot=true2024.0327.55
LFT-Hassle-free-ALPS-TMConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0328
Block (Ma, Fang, and Wang 2023)Pruning Ratio=20%, Zero-shot=true2024.0328.08
LFT-3BASiL-TMConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0329
LFT-OATSConfig=2:4+64LR, Fine-Tuning=LoRA2026.0329.6
LFT-3BASiLConfig=3:8+64LR, Fine-Tuning=LoRA2026.0329.74
LFT-3BASiL-TMConfig=3:8+64LR, Fine-Tuning=LoRA2026.0329.94
LFT-3BASiLConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0330.05
LFT-Hassle-free-ALPS-TMConfig=3:8+64LR, Fine-Tuning=LoRA2026.0330.32
LFT-Hassle-free-SparseGPT-TMConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0331.65
LFT-Hassle-free-ALPSConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0331.99
LFT-Hassle-free-ALPSConfig=3:8+64LR, Fine-Tuning=LoRA2026.0332.26
LFT-Hassle-free-SparseGPT-TMConfig=3:8+64LR, Fine-Tuning=LoRA2026.0332.36
LFT-OATSConfig=4:8+64LR, Fine-Tuning=LoRA2026.0332.9
LFT-Hassle-free-SparseGPTConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0333.94
LFT-Hassle-free-SparseGPTConfig=3:8+64LR, Fine-Tuning=LoRA2026.0334.55
LFT-OATS-TMConfig=3:8+64LR, Fine-Tuning=LoRA2026.0334.83
LFT-OATS-TMConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0336.25
L2 (Ma, Fang, and Wang 2023)Pruning Ratio=20%, Zero-shot=true2024.0338.05
LFT-OATSConfig=3:8+64LR, Fine-Tuning=LoRA2026.0341.74
LFT-3BASiLConfig=2:8+64LR, Fine-Tuning=LoRA2026.0345.13
LFT-3BASiL-TMConfig=2:8+64LR, Fine-Tuning=LoRA2026.0347.11
LFT-OATSConfig=2:8+64LR, Fine-tuning Protocol=LFT, Base Model=Meta-Llama-3-8B2026.0349.71
LFT-Hassle-free-ALPS-TMConfig=2:8+64LR, Fine-Tuning=LoRA2026.0350.15
LFT-Hassle-free-SparseGPT-TMConfig=2:8+64LR, Fine-Tuning=LoRA2026.0355.53
LFT-Hassle-free-ALPSConfig=2:8+64LR, Fine-Tuning=LoRA2026.0357.45
LFT-Hassle-free-SparseGPTConfig=2:8+64LR, Fine-Tuning=LoRA2026.0363.87
LFT-OATS-TMConfig=2:8+64LR, Fine-Tuning=LoRA2026.0368.98
LFT-OATSConfig=2:8+64LR, Fine-Tuning=LoRA2026.0395.05