Common-sense reasoning on BoolQ In-Distribution
90.5AccuracyENS
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ENSBackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 90.5 | 7.35 | 0.42 | |
| MAPBackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 90.09 | 8.4 | 0.52 | |
| MCDBackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 89.93 | 7.95 | 0.49 | |
| MLEBackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 89.58 | 8.69 | 0.52 | |
| DALorRABackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 89.43 | 1.82 | 0.27 | |
| LABackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 89.36 | 1.93 | 0.43 | |
| PoLAR-MLEBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 89.16 | 2.12 | 0.27 | |
| PoLAR-LABackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 89.16 | 1.88 | 0.27 | |
| PoLAR-LA-LLBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 89.16 | 1.96 | 0.27 | |
| PoLAR-VBLL (w/o LA)Backbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 88.88 | 2.14 | 0.32 | |
| PoLAR-VBLLBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 88.88 | 1.77 | 0.26 | |
| TFBBackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 88.84 | 3.8 | 0.29 | |
| BLoBBackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 88.69 | 2.46 | 0.27 | |
| C-LoRABackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 88.26 | 3.63 | 0.29 | |
| MLEBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 87.86 | 4.76 | 0.27 | |
| LABackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 87.86 | 2.33 | 0.27 | |
| TFBBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 87.83 | 4.21 | 0.27 | |
| C-LoRABackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 87.77 | 4.36 | 0.3 | |
| PoLAR-BLoBBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 87.73 | 2.38 | 0.29 | |
| BLoBBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 87.54 | 2.61 | 0.31 | |
| TFB-LLBackbone=Llama-3.1-8B, Fine-tuning steps=50002026.04 | 87.53 | 3.13 | 0.27 |