List your company

Explaining LLMs on NQ

11.76CRRProxy Explanation Framework
2.02564.55287.089.6072May 18, 2025

Evaluation results

MethodLinks
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=LIME, CRR Type=local2025.0511.76
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=Kernel SHAP, CRR Type=local2025.0511.76
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=Kernel SHAP, CRR Type=local2025.0511.11
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=LIME, CRR Type=local2025.0510.53
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=Kernel SHAP, CRR Type=local2025.0510.53
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=Kernel SHAP, CRR Type=local2025.0510.53
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=Kernel SHAP, CRR Type=local2025.0510
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=LIME, CRR Type=local2025.059.09
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=Kernel SHAP, CRR Type=local2025.059.09
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=LIME, CRR Type=local2025.059.09
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=LIME, CRR Type=local2025.058.33
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=Kernel SHAP, CRR Type=local2025.058.33
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=Kernel SHAP, CRR Type=max2025.058.2
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=LIME, CRR Type=max2025.057.41
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=Kernel SHAP, CRR Type=max2025.057.41
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=Kernel SHAP, CRR Type=mean2025.057.29
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=Kernel SHAP, CRR Type=max2025.057.14
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=LIME, CRR Type=local2025.056.9
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=LIME, CRR Type=max2025.056.67
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=LIME, CRR Type=local2025.056.67
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=Kernel SHAP, CRR Type=max2025.056.67
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=Kernel SHAP, CRR Type=max2025.056.67
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=Kernel SHAP, CRR Type=max2025.056.25
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=LIME, CRR Type=max2025.056.25
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=LIME, CRR Type=max2025.056.06
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=LIME, CRR Type=max2025.056.06
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=Kernel SHAP, CRR Type=max2025.056.06
Proxy Explanation FrameworkTarget Model=GPT-4o, Explanation Technique=LIME, CRR Type=mean2025.055.7
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.055.13
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=LIME, CRR Type=max2025.055.13
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.054.92
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=LIME, CRR Type=max2025.054.17
Proxy Explanation FrameworkTarget Model=Qwen 2.5 72B, Explanation Technique=LIME, CRR Type=mean2025.053.97
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.053.86
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=Kernel SHAP, CRR Type=mean2025.053.85
Proxy Explanation FrameworkTarget Model=Qwen 2.5 32B, Explanation Technique=LIME, CRR Type=mean2025.053.74
Proxy Explanation FrameworkTarget Model=Llama 3.1 70B, Explanation Technique=LIME, CRR Type=mean2025.053.72
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=LIME, CRR Type=mean2025.053.65
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=Kernel SHAP, CRR Type=mean2025.053.49
Proxy Explanation FrameworkTarget Model=GPT-4o mini, Explanation Technique=Kernel SHAP, CRR Type=mean2025.053.43
Proxy Explanation FrameworkTarget Model=Qwen 2.5 14B, Explanation Technique=LIME, CRR Type=mean2025.052.72
Proxy Explanation FrameworkTarget Model=DeepSeek V3, Explanation Technique=LIME, CRR Type=mean2025.052.4