Question Answering on ARC Easy (ARC-E)
85.4Accuracy (ARC-E)MARI
Evaluation Results
| Method | Links | |
|---|---|---|
| MARIModel=Qwen2.5-14B2026.05 | 85.4 | |
| CAAModel=Qwen2.5-14B2026.05 | 85 | |
| NL-ITIModel=Qwen2.5-14B2026.05 | 84.8 | |
| VanillaModel=Qwen2.5-14B2026.05 | 84.7 | |
| ReFTModel=Qwen2.5-14B2026.05 | 84.6 | |
| ITIModel=Qwen2.5-14B2026.05 | 84.5 | |
| MARIModel=Qwen2.5-32B2026.05 | 78.5 | |
| CAAModel=Qwen2.5-32B2026.05 | 78.1 | |
| MARIModel=Llama-3-8B2026.05 | 78 | |
| NL-ITIModel=Qwen2.5-32B2026.05 | 77.9 | |
| VanillaModel=Qwen2.5-32B2026.05 | 77.8 | |
| CAAModel=Llama-3-8B2026.05 | 77.7 | |
| ReFTModel=Qwen2.5-32B2026.05 | 77.7 | |
| ITIModel=Qwen2.5-32B2026.05 | 77.6 | |
| NL-ITIModel=Llama-3-8B2026.05 | 77.4 | |
| VanillaModel=Llama-3-8B2026.05 | 77.3 | |
| ITIModel=Llama-3-8B2026.05 | 77.2 | |
| ReFTModel=Llama-3-8B2026.05 | 77.1 | |
| CAAModel=Qwen2-7B2026.05 | 75.8 | |
| NL-ITIModel=Qwen2-7B2026.05 | 75.6 | |
| VanillaModel=Qwen2-7B2026.05 | 75.5 | |
| ReFTModel=Qwen2-7B2026.05 | 75.4 | |
| ITIModel=Qwen2-7B2026.05 | 75.3 | |
| MARIModel=Qwen2-7B2026.05 | 75.2 | |
| CAAModel=Llama-2-13B2026.05 | 53.9 | |
| NL-ITIModel=Llama-2-13B2026.05 | 53.8 | |
| VanillaModel=Llama-2-13B2026.05 | 53.7 | |
| ReFTModel=Llama-2-13B2026.05 | 53.6 | |
| ITIModel=Llama-2-13B2026.05 | 53.5 | |
| MARIModel=Llama-2-13B2026.05 | 53.4 | |
| MARIModel=Llama-2-7B2026.05 | 52.6 | |
| CAAModel=Llama-2-7B2026.05 | 52.2 | |
| NL-ITIModel=Llama-2-7B2026.05 | 52 | |
| VanillaModel=Llama-2-7B2026.05 | 51.9 | |
| ReFTModel=Llama-2-7B2026.05 | 51.8 | |
| ITIModel=Llama-2-7B2026.05 | 51.7 |