List your company

Explaining LLMs on MMLU

6.31CRRProxy Explanation Framework
1.64042.85274.0655.2773May 18, 2025

Evaluation results

MethodLinks
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=Kernel SHAP, CRR Type=local2025.056.31
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=Kernel SHAP, CRR Type=local2025.056.1
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=Kernel SHAP, CRR Type=local2025.056.1
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=LIME, CRR Type=local2025.056.07
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=Kernel SHAP, CRR Type=local2025.055.78
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=LIME, CRR Type=local2025.055.77
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=Kernel SHAP, CRR Type=local2025.055.77
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=LIME, CRR Type=local2025.055.64
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=LIME, CRR Type=local2025.055.62
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=Kernel SHAP, CRR Type=local2025.055.62
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=LIME, CRR Type=local2025.055.32
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=Kernel SHAP, CRR Type=local2025.055.31
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=LIME, CRR Type=local2025.055.3
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=LIME, CRR Type=local2025.054.98
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=LIME, CRR Type=max2025.054.84
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=Kernel SHAP, CRR Type=max2025.054.84
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=LIME, CRR Type=max2025.053.96
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=Kernel SHAP, CRR Type=max2025.053.96
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=Kernel SHAP, CRR Type=max2025.053.66
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=LIME, CRR Type=mean2025.053.41
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=Kernel SHAP, CRR Type=mean2025.053.41
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=LIME, CRR Type=max2025.053.4
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=Kernel SHAP, CRR Type=max2025.053.2
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=Kernel SHAP, CRR Type=max2025.053.19
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=Kernel SHAP, CRR Type=max2025.053.16
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=LIME, CRR Type=max2025.053.15
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=LIME, CRR Type=max2025.053.05
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=Kernel SHAP, CRR Type=max2025.053.05
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=LIME, CRR Type=max2025.052.99
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.052.93
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=LIME, CRR Type=mean2025.052.92
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=LIME, CRR Type=max2025.052.88
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.052.76
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=LIME, CRR Type=mean2025.052.47
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=Kernel SHAP, CRR Type=mean2025.052.27
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.052.16
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=LIME, CRR Type=mean2025.052.15
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=LIME, CRR Type=mean2025.052.06
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=Kernel SHAP, CRR Type=mean2025.051.96
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.051.9
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=LIME, CRR Type=mean2025.051.83
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=LIME, CRR Type=mean2025.051.82