Science Question Answering on AI2SciE (test)
95.8AccuracyFew-shot*
Evaluation Results
| Method | Links | |
|---|---|---|
| Few-shot*Backbone=Qwen2.5-7B, Few-shot settings=3-shot OOD2025.09 | 95.8 | |
| ICRBackbone=Qwen2.5-7B2025.09 | 93.2 | |
| M²IVBackbone=Qwen2.5-7B2025.09 | 92.4 | |
| I2CLBackbone=Qwen2.5-7B2025.09 | 91.8 | |
| LIVEBackbone=Qwen2.5-7B2025.09 | 91.8 | |
| Zero-shotBackbone=Qwen2.5-7B2025.09 | 90.4 | |
| Few-shot*Backbone=Llama2-7B, Few-shot settings=3-shot OOD2025.09 | 45.4 | |
| ICRBackbone=Llama2-7B2025.09 | 37.2 | |
| M²IVBackbone=Llama2-7B2025.09 | 35.2 | |
| I2CLBackbone=Llama2-7B2025.09 | 35 | |
| Zero-shotBackbone=Llama2-7B2025.09 | 34.8 | |
| LIVEBackbone=Llama2-7B2025.09 | 34.6 |