List your company

Zero-shot Evaluation on ARC-C, ARC-E, HellaSwag, PIQA, and WinoGrande (test val)

76Average AccuracyFP16 Baseline
39.39248.89658.467.904May 18, 2025Jun 20, 2025Jul 23, 2025Aug 25, 2025Sep 27, 2025Oct 30, 2025Dec 3, 2025

Evaluation results

MethodLinks
FP16 BaselinePruning Ratio=0%, Backbone=Nemotron-H-8B, Fine-tuned=false2025.1276—————
FP16 BaselinePruning Ratio=0%, Backbone=Bamba-v2-9B, Fine-tuned=false2025.1274.6—————
FP16 BaselinePruning Ratio=0%, Backbone=Llama-3.1-8B, Fine-tuned=false2025.1274—————
UniQLPruning Ratio=15%, Backbone=Nemotron-H-8B, Fine-tuned=true2025.1273—————
UniQLPruning Ratio=15%, Backbone=Bamba-v2-9B, Fine-tuned=true2025.1272.9—————
FP16 BaselinePruning Ratio=0%, Backbone=Qwen-2.5-7B, Fine-tuned=false2025.1272.4—————
MoDeGPTPruning Ratio=15%, Backbone=Llama-3.1-8B, Fine-tuned=false2025.1272.4—————
UniQLPruning Ratio=15%, Backbone=Llama-3.1-8B, Fine-tuned=true2025.1271.9—————
UniQLPruning Ratio=15%, Backbone=Bamba-v2-9B, Fine-tuned=false2025.1270.9—————
FP16 BaselinePruning Ratio=0%, Backbone=Mamba2-8B, Fine-tuned=false2025.1270.6—————
UniQLPruning Ratio=15%, Backbone=Llama-3.1-8B, Fine-tuned=false2025.1270.5—————
UniQLPruning Ratio=15%, Backbone=Qwen-2.5-7B, Fine-tuned=true2025.1270—————
UniQLPruning Ratio=25%, Backbone=Bamba-v2-9B, Fine-tuned=true2025.1269.7—————
UniQLPruning Ratio=25%, Backbone=Llama-3.1-8B, Fine-tuned=true2025.1269.6—————
SVD-LLMPruning Ratio=15%, Backbone=Qwen-2.5-7B, Fine-tuned=true2025.1269.5—————
UniQLPruning Ratio=15%, Backbone=Qwen-2.5-7B, Fine-tuned=false2025.1269.1—————
UniQLPruning Ratio=15%, Backbone=Nemotron-H-8B, Fine-tuned=false2025.1268.9—————
FP16 BaselinePruning Ratio=0%, Backbone=Llama-2-7B, Fine-tuned=false2025.1268.8—————
DenseSparsity=0%, Backbone=Mistral-7B2025.0568.7248.979.660.980.373.9
UniQLPruning Ratio=25%, Backbone=Nemotron-H-8B, Fine-tuned=true2025.1267.3—————
UniQLPruning Ratio=15%, Backbone=Llama-2-7B, Fine-tuned=true2025.1267.2—————
UniQLPruning Ratio=25%, Backbone=Llama-3.1-8B, Fine-tuned=false2025.1267—————
SVD-LLMPruning Ratio=25%, Backbone=Qwen-2.5-7B, Fine-tuned=true2025.1266.8—————
UniQLPruning Ratio=15%, Backbone=Llama-2-7B, Fine-tuned=false2025.1266.7—————
UniQLPruning Ratio=15%, Backbone=Mamba2-8B, Fine-tuned=true2025.1266.4—————
UniQLPruning Ratio=25%, Backbone=Bamba-v2-9B, Fine-tuned=false2025.1266.4—————
MoDeGPTPruning Ratio=15%, Backbone=Llama-2-7B, Fine-tuned=false2025.1266.2—————
UniQLPruning Ratio=25%, Backbone=Qwen-2.5-7B, Fine-tuned=true2025.1265.8—————
EvoPressSparsity=50%, Latency=122m, Zero-shot=true, Backbone=Mistral-7B2025.0565.7444.8876.8556.4679.1671.35
UniQLPruning Ratio=15%, Backbone=Mamba2-8B, Fine-tuned=false2025.1265.6—————
UniCuCoSparsity=50%, Latency=<1s, Zero-shot=true, Backbone=Mistral-7B2025.0565.2344.1176.7355.6678.471.27
UniformSparsity=50%, Latency=<1s, Zero-shot=true, Backbone=Mistral-7B2025.0565.143.776.755.778.471
OWLSparsity=50%, Latency=40m, Zero-shot=true, Backbone=Mistral-7B2025.056543.976.955.478.570.3
MoDeGPTPruning Ratio=25%, Backbone=Llama-3.1-8B, Fine-tuned=false2025.1264.9—————
UniQLPruning Ratio=25%, Backbone=Llama-2-7B, Fine-tuned=true2025.1264.9—————
SVD-LLMPruning Ratio=15%, Backbone=Llama-2-7B, Fine-tuned=true2025.1264.7—————
SVD-LLMPruning Ratio=15%, Backbone=Llama-3.1-8B, Fine-tuned=true2025.1264.5—————
UniQLPruning Ratio=25%, Backbone=Llama-2-7B, Fine-tuned=false2025.1263.7—————
MoDeGPTPruning Ratio=25%, Backbone=Llama-2-7B, Fine-tuned=false2025.1263.4—————
UniQLPruning Ratio=25%, Backbone=Mamba2-8B, Fine-tuned=true2025.1262.7—————
SVD-LLMPruning Ratio=15%, Backbone=Qwen-2.5-7B, Fine-tuned=false2025.1262.6—————
SVD-LLMPruning Ratio=25%, Backbone=Llama-2-7B, Fine-tuned=true2025.1262.4—————
UniQLPruning Ratio=25%, Backbone=Qwen-2.5-7B, Fine-tuned=false2025.1262.1—————
UniCuCoSparsity=60%, Latency=<1s, Zero-shot=true, Backbone=Mistral-7B2025.0561.5739.3372.949.9175.7969.93
EvoPressSparsity=60%, Latency=121m, Zero-shot=true, Backbone=Mistral-7B2025.0561.138.0572.5649.9176.0168.98
UniformSparsity=60%, Latency=<1s, Zero-shot=true, Backbone=Mistral-7B2025.0560.823872.449.47569.3
UniQLPruning Ratio=25%, Backbone=Nemotron-H-8B, Fine-tuned=false2025.1260.6—————
OWLSparsity=60%, Latency=40m, Zero-shot=true, Backbone=Mistral-7B2025.0560.5238.571.946.975.170.2
UniQLPruning Ratio=25%, Backbone=Mamba2-8B, Fine-tuned=false2025.1259.8—————
SVD-LLMPruning Ratio=25%, Backbone=Llama-3.1-8B, Fine-tuned=true2025.1259.5—————
SVD-LLMPruning Ratio=15%, Backbone=Llama-3.1-8B, Fine-tuned=false2025.1256.7—————
EvoPressSparsity=70%, Latency=120m, Zero-shot=true, Backbone=Mistral-7B2025.0556.4833.4567.1343.9172.6365.27
SVD-LLMPruning Ratio=15%, Backbone=Llama-2-7B, Fine-tuned=false2025.1256.3—————
SVD-LLMPruning Ratio=25%, Backbone=Qwen-2.5-7B, Fine-tuned=false2025.1253.2—————
UniCuCoSparsity=70%, Latency=<1s, Zero-shot=true, Backbone=Mistral-7B2025.0552.8729.164.2739.0569.0462.9
MoDeGPTPruning Ratio=15%, Backbone=Qwen-2.5-7B, Fine-tuned=false2025.1252.1—————
OWLSparsity=70%, Latency=40m, Zero-shot=true, Backbone=Mistral-7B2025.0551.9227.962.638.66763.5
SVD-LLMPruning Ratio=25%, Backbone=Llama-2-7B, Fine-tuned=false2025.1250.8—————
UniformSparsity=70%, Latency=<1s, Zero-shot=true, Backbone=Mistral-7B2025.0549.8827.160.936.165.959.4
SVD-LLMPruning Ratio=25%, Backbone=Llama-3.1-8B, Fine-tuned=false2025.1245.8—————
MoDeGPTPruning Ratio=25%, Backbone=Qwen-2.5-7B, Fine-tuned=false2025.1240.8—————