Commonsense Question Answering on CommonSenseQA
88.9AccuracyHUMAN
Evaluation Results
| Method | Links | |
|---|---|---|
| HUMAN2021.03 | 88.9 | |
| EPR-KMeansModel=Qwen2.5-72B, Shots=1282025.06 | 87.96 | |
| BGE-KMeansModel=Qwen2.5-72B, Shots=1282025.06 | 87.71 | |
| CLGModel=Qwen2.5-72B, Shots=1282025.06 | 87.71 | |
| Best-of-NModel=Qwen2.5-72B, Shots=1282025.06 | 87.55 | |
| RandomModel=Qwen2.5-72B, Shots=1282025.06 | 87.47 | |
| BM25-MajorModel=Qwen2.5-72B, Shots=1282025.06 | 86.65 | |
| Llama 3 405BParameters=405B2024.07 | 85.8 | |
| Latent-BayesianModel=Qwen2.5-72B, Shots=1282025.06 | 85.75 | |
| QWEN-3-30B MOEPARAMS=30B, Variant=FCSD2025.11 | 84.9 | |
| QWEN-3-30B MOEPARAMS=30B, Variant=ORIGINAL2025.11 | 84.8 | |
| RandomModel=Llama3-70B, Shots=1282025.06 | 84.73 | |
| Latent-BayesianModel=Llama3-70B, Shots=1282025.06 | 84.6 | |
| CLGModel=Llama3-70B, Shots=1282025.06 | 84.52 | |
| LLAMA-4-SCOUTPARAMS=109B, Variant=ORIGINAL2025.11 | 84.4 | |
| Best-of-NModel=Llama3-70B, Shots=1282025.06 | 84.19 | |
| Llama 3 70BParameters=70B2024.07 | 84.1 | |
| BGE-KMeansModel=Llama3-70B, Shots=1282025.06 | 83.78 | |
| EPR-KMeansModel=Llama3-70B, Shots=1282025.06 | 83.62 | |
| Mixtral 8×22BParameters=8×22B2024.07 | 82.4 | |
| LLAMA-4-SCOUTPARAMS=109B, Variant=FCSD2025.11 | 82.4 | |
| TAIAModel=Qwen1.5-32B, Infer Mode=TAIA2024.05 | 82.31 | |
| BM25-MajorModel=Llama3-70B, Shots=1282025.06 | 82.23 | |
| BaseModel=Qwen1.5-32B, Infer Mode=Base2024.05 | 80.92 | |
| UNICORN2021.03 | 79.3 | |
| UNIFIEDQA-11B2021.03 | 79.1 | |
| SDPO2026.02 | 78.71 | |
| Qwen3-8B2026.02 | 78.46 | |
| BaseModel=Qwen1.5-7B, Infer Mode=Base2024.05 | 78.3 | |
| T5-11B2021.03 | 78.1 | |
| LoRAModel=Qwen1.5-32B, Infer Mode=LoRA2024.05 | 77.81 | |
| BaseModel=Qwen1.5-14B, Infer Mode=Base2024.05 | 77.72 | |
| LoRAModel=Qwen1.5-7B, Infer Mode=LoRA2024.05 | 77.31 | |
| TAIAModel=Qwen1.5-7B, Infer Mode=TAIA2024.05 | 77.31 | |
| TAIAModel=Qwen1.5-14B, Infer Mode=TAIA2024.05 | 77.31 | |
| Llama 3 8BParameters=8B2024.07 | 75 | |
| LoRAModel=Qwen1.5-14B, Infer Mode=LoRA2024.05 | 74.45 | |
| Gemma 7BParameters=7B2024.07 | 74.4 | |
| ROBERTA-LARGE2021.03 | 72.1 | |
| LLAMA-4-SCOUTPARAMS=109B, Variant=SFT2025.11 | 72 | |
| Mistral 7BParameters=7B2024.07 | 71.2 | |
| DEEPSEEK-V2-LITEPARAMS=16B, Variant=FCSD2025.11 | 70.1 | |
| DEEPSEEK-V2-LITEPARAMS=16B, Variant=ORIGINAL2025.11 | 69.1 | |
| QWEN-3-30B MOEPARAMS=30B, Variant=SFT2025.11 | 68.9 | |
| DEEPSEEK-V2-LITEPARAMS=16B, Variant=SFT2025.11 | 65.5 | |
| OLTQAConfiguration=Full2023.05 | 57.08 | |
| OLTQAAblation=without Pm2023.05 | 56.91 | |
| OLTQAKnowledge Distillation=Back KD2023.05 | 56.73 | |
| OLTQAKnowledge Distillation=Static MKD2023.05 | 56.1 | |
| OLTQAKnowledge Distillation=without MKD2023.05 | 55.61 | |
| OLTQARetriever=EPR2023.05 | 54.95 | |
| Muppet2023.05 | 53.79 | |
| Hyperformer++2023.05 | 53.24 | |
| OLTQARe-ranker=excluded2023.05 | 52.83 | |
| OLTQARetriever=BM252023.05 | 52.66 | |
| OLTQAAblation=without Pk2023.05 | 52.5 | |
| UnifiedQA2023.05 | 51.92 | |
| ProQA2023.05 | 51.52 | |
| EPR2023.05 | 45.3 | |
| LTVPrompt=Zero-shot2025.02 | 41 | |
| + LTVprompt=zero-shot2025.02 | 41 | |
| Pre-editedBackbone=LLaMA2, Number of Edits=8000, Edit Dataset=zsRE2026.01 | 32.84 | |
| HORSEBackbone=LLaMA2, Number of Edits=8000, Edit Dataset=zsRE2026.01 | 29.87 | |
| MALMENBackbone=LLaMA2, Number of Edits=8000, Edit Dataset=zsRE2026.01 | 29.81 | |
| ICVPrompt=Zero-shot2025.02 | 28.9 | |
| + FVprompt=zero-shot2025.02 | 28.9 | |
| PMETBackbone=LLaMA2, Number of Edits=8000, Edit Dataset=zsRE2026.01 | 25.63 | |
| LoRAPrompt=Zero-shot2025.02 | 24.8 | |
| FVPrompt=Zero-shot2025.02 | 23 | |
| Transformerprompt=zero-shot2025.02 | 23 | |
| TVPrompt=Zero-shot2025.02 | 22 | |
| + SVprompt=zero-shot2025.02 | 22 | |
| SVPrompt=Zero-shot2025.02 | 21.7 | |
| + ICVprompt=zero-shot2025.02 | 21.7 | |
| LTVPrompt=5-shot2025.02 | 21.1 | |
| + TVprompt=zero-shot2025.02 | 21.1 | |
| + LTVprompt=few-shot2025.02 | 21.1 | |
| FVPrompt=5-shot2025.02 | 20.3 | |
| SVPrompt=5-shot2025.02 | 20.3 | |
| Transformerprompt=few-shot2025.02 | 20.3 | |
| + ICVprompt=few-shot2025.02 | 20.3 | |
| MEMITBackbone=LLaMA2, Number of Edits=8000, Edit Dataset=zsRE2026.01 | 20.23 | |
| ICVPrompt=5-shot2025.02 | 20 | |
| + FVprompt=few-shot2025.02 | 20 | |
| MENDBackbone=LLaMA2, Number of Edits=8000, Edit Dataset=zsRE2026.01 | 19.74 | |
| TVPrompt=5-shot2025.02 | 19.5 | |
| + SVprompt=few-shot2025.02 | 19.5 | |
| LoRAPrompt=5-shot2025.02 | 18.6 | |
| TransformerPrompt=5-shot2025.02 | 18 | |
| Standard Promptingprompt=few-shot2025.02 | 18 | |
| TransformerPrompt=Zero-shot2025.02 | 10.2 | |
| Standard Promptingprompt=zero-shot2025.02 | 10.2 |