Reading Comprehension on DROP MRQA out-of-domain
0.4884EMPrompt tuning
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Prompt tuningBackbone=Llama-3.2-3B-Base, Context=with context2026.02 | 0.4884 | 0.5718 | |
| ComprExITBackbone=Llama-3.2-3B-Base2026.02 | 0.4358 | 0.5316 | |
| BeaconBackbone=Llama-3.2-3B-Base2026.02 | 0.3939 | 0.5055 | |
| ICAEBackbone=Llama-3.2-3B-Base2026.02 | 0.3533 | 0.4489 | |
| Prompt tuningBackbone=Llama-3.2-1B-Base, Context=with context2026.02 | 0.3254 | 0.4314 | |
| 500xBackbone=Llama-3.2-3B-Base2026.02 | 0.2974 | 0.3912 | |
| ComprExITBackbone=Llama-3.2-1B-Base2026.02 | 0.2901 | 0.3986 | |
| ICAEBackbone=Llama-3.2-1B-Base2026.02 | 0.2415 | 0.3323 | |
| Zero-shotBackbone=Llama-3.2-3B-Base, Context=with context2026.02 | 0.2369 | 0.3745 | |
| BeaconBackbone=Llama-3.2-1B-Base2026.02 | 0.2182 | 0.2924 | |
| Zero-shotBackbone=Llama-3.2-1B-Base, Context=with context2026.02 | 0.1843 | 0.323 | |
| Zero-shotBackbone=Llama-3.2-3B-Base, Context=without context2026.02 | 0.1617 | 0.2126 | |
| Zero-shotBackbone=Llama-3.2-1B-Base, Context=without context2026.02 | 0.1484 | 0.1973 | |
| 500xBackbone=Llama-3.2-1B-Base2026.02 | 0.0818 | 0.1832 |