List your company

Accuracy on Winogrande (Common-sense Reasoning)

85.64AccuracyST-D
49.926459.198268.4777.7418Nov 6, 2025Dec 15, 2025Jan 24, 2026Mar 4, 2026Apr 13, 2026May 22, 2026Jul 1, 2026

Evaluation results

MethodLinks
ST-DBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.0685.64
ST-SBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.0685.16
BaseBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.0684.93
Star1Backbone Model=DeepSeek-R1-Distill-Llama-70B2026.0684.61
SafePathBackbone Model=DeepSeek-R1-Distill-Llama-70B2026.0684.53
BaseBackbone Model=Qwen3-8B2026.0678.22
ST-SBackbone Model=Qwen3-8B2026.0677.82
ST-DBackbone Model=Qwen3-8B2026.0677.74
VanillaBackbone=Mixtral-8x7B-v0.1, Sparsity=0%2025.1177.2
BaseBackbone Model=Qwen3-32B2026.0676.48
ST-DBackbone Model=Qwen3-32B2026.0676.48
FP16BIT=16, Evaluation Protocol=Few-shot2026.0776.48
PuzzleMoEBackbone=Mixtral-8x7B-v0.1, Sparsity=50%2025.1175.8
PuzzleMoEBackbone=Mixtral-8x7B-v0.1, Sparsity=25%2025.1175.5
HC-SMoEBackbone=Mixtral-8x7B-v0.1, Sparsity=25%2025.1175.4
ST-SBackbone Model=Qwen3-32B2026.0675.14
NAEEBackbone=Mixtral-8x7B-v0.1, Sparsity=25%2025.1175.1
D2Backbone=Mixtral-8x7B-v0.1, Sparsity=20%2025.1175
Sub-MoEBackbone=Mixtral-8x7B-v0.1, Sparsity=25%2025.1175
WandaBackbone=Mixtral-8x7B-v0.1, Sparsity=2:42025.1174.5
Star1Backbone Model=Qwen3-32B2026.0674.03
GSRQBIT=2, Evaluation Protocol=Few-shot2026.0773.16
NAEEBackbone=Mixtral-8x7B-v0.1, Sparsity=50%2025.1173
D2Backbone=Mixtral-8x7B-v0.1, Sparsity=40%2025.1173
STUNBackbone=Mixtral-8x7B-v0.1, Sparsity=25%2025.1172.7
SafePathBackbone Model=Qwen3-32B2026.0672.45
HC-SMoEBackbone=Mixtral-8x7B-v0.1, Sparsity=50%2025.1172.1
Sub-MoEBackbone=Mixtral-8x7B-v0.1, Sparsity=50%2025.1172
PuzzleMoEBackbone=Deepseek-MoE-16b, Sparsity=25%2025.1170.6
WandaBackbone=Qwen3-MoE-30B-A3B, Sparsity=2:42025.1170.5
VanillaBackbone=Qwen3-MoE-30B-A3B, Sparsity=0%2025.1170.4
PuzzleMoEBackbone=Qwen3-MoE-30B-A3B, Sparsity=25%2025.1170.4
PuzzleMoEBackbone=Deepseek-MoE-16b, Sparsity=50%2025.1170.3
VanillaBackbone=Deepseek-MoE-16b, Sparsity=0%2025.1170.1
PuzzleMoEBackbone=Qwen3-MoE-30B-A3B, Sparsity=50%2025.1170.1
Sub-MoEBackbone=Deepseek-MoE-16b, Sparsity=25%2025.1170
PuzzleMoEBackbone=Qwen1.5-MoE-A2.7B, Sparsity=25%2025.1169.6
PuzzleMoEBackbone=Qwen1.5-MoE-A2.7B, Sparsity=50%2025.1169.4
VQLLMBIT=2, Evaluation Protocol=Few-shot2026.0769.3
WandaBackbone=Deepseek-MoE-16b, Sparsity=2:42025.1169.2
D2Backbone=Qwen3-MoE-30B-A3B, Sparsity=20%2025.1169.1
D2Backbone=Deepseek-MoE-16b, Sparsity=20%2025.1169
GSRQBIT=1, Evaluation Protocol=Few-shot2026.0768.98
VanillaBackbone=Qwen1.5-MoE-A2.7B, Sparsity=0%2025.1168.9
ST-SBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.0668.35
WandaBackbone=Qwen1.5-MoE-A2.7B, Sparsity=2:42025.1167.8
ST-DBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.0667.32
D2Backbone=Qwen1.5-MoE-A2.7B, Sparsity=20%2025.1166.8
BaseBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.0666.38
HC-SMoEBackbone=Qwen1.5-MoE-A2.7B, Sparsity=25%2025.1166.1
Sub-MoEBackbone=Qwen3-MoE-30B-A3B, Sparsity=25%2025.1166
HC-SMoEBackbone=Deepseek-MoE-16b, Sparsity=25%2025.1165.7
Star1Backbone Model=Qwen3-8B2026.0665.51
SafePathBackbone Model=Qwen3-8B2026.0665.43
D2Backbone=Deepseek-MoE-16b, Sparsity=40%2025.1165
Sub-MoEBackbone=Deepseek-MoE-16b, Sparsity=50%2025.1165
GSRQBIT=0.75, Evaluation Protocol=Few-shot2026.0763.54
HC-SMoEBackbone=Qwen3-MoE-30B-A3B, Sparsity=25%2025.1163.2
Sub-MoEBackbone=Qwen3-MoE-30B-A3B, Sparsity=50%2025.1163
D2Backbone=Qwen1.5-MoE-A2.7B, Sparsity=40%2025.1160.3
D2Backbone=Qwen3-MoE-30B-A3B, Sparsity=40%2025.1160.2
HC-SMoEBackbone=Qwen1.5-MoE-A2.7B, Sparsity=50%2025.1156.1
HC-SMoEBackbone=Deepseek-MoE-16b, Sparsity=50%2025.1155.3
HC-SMoEBackbone=Qwen3-MoE-30B-A3B, Sparsity=50%2025.1154
Star1Backbone Model=DeepSeek-R1-Distill-Llama-8B2026.0653.59
VQLLMBIT=1, Evaluation Protocol=Few-shot2026.0753.12
SafePathBackbone Model=DeepSeek-R1-Distill-Llama-8B2026.0651.3