Reading Comprehension on DuoRC MRQA out-of-domain
45.5EMPrompt tuning
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Prompt tuningBackbone=Llama-3.2-3B-Base, Context=with context2026.02 | 45.5 | 55.63 | |
| Prompt tuningBackbone=Llama-3.2-1B-Base, Context=with context2026.02 | 38.51 | 47.4 | |
| ComprExITBackbone=Llama-3.2-3B-Base2026.02 | 34.31 | 44.98 | |
| ComprExITBackbone=Llama-3.2-1B-Base2026.02 | 28.78 | 38.77 | |
| Zero-shotBackbone=Llama-3.2-3B-Base, Context=with context2026.02 | 28.31 | 42.36 | |
| ICAEBackbone=Llama-3.2-3B-Base2026.02 | 23.65 | 32.81 | |
| 500xBackbone=Llama-3.2-3B-Base2026.02 | 22.25 | 30.91 | |
| Zero-shotBackbone=Llama-3.2-1B-Base, Context=with context2026.02 | 17.46 | 30.54 | |
| ICAEBackbone=Llama-3.2-1B-Base2026.02 | 12.59 | 18.8 | |
| BeaconBackbone=Llama-3.2-3B-Base2026.02 | 4.73 | 7.95 | |
| BeaconBackbone=Llama-3.2-1B-Base2026.02 | 2.4 | 4.9 | |
| Zero-shotBackbone=Llama-3.2-3B-Base, Context=without context2026.02 | 2.07 | 4.44 | |
| 500xBackbone=Llama-3.2-1B-Base2026.02 | 1.07 | 3.91 | |
| Zero-shotBackbone=Llama-3.2-1B-Base, Context=without context2026.02 | 0.6 | 2.55 |