Question Answering on EntityQuestions
69.5AccuracyRA
Evaluation results
| Method | Links | |
|---|---|---|
| RATraining Method=Fine-tuning, Injection Method=RA2023.05 | 69.5 | |
| TOME-22021.10 | 66 | |
| TOME-12021.10 | 62.1 | |
| REALM2021.10 | 59 | |
| Map-tuningTraining Method=Adapter, Injection Method=Map-tuning2023.05 | 50.8 | |
| Map-tuningTraining Method=LoRA, Injection Method=Map-tuning2023.05 | 49.6 | |
| Map-tuningTraining Method=Fine-tuning, Injection Method=Map-tuning2023.05 | 49 | |
| Map-tuningTraining Method=BitFit, Injection Method=Map-tuning2023.05 | 48.4 | |
| BERT-baseTraining Method=Adapter, Injection Method=None2023.05 | 43.6 | |
| BERT-baseTraining Method=LoRA, Injection Method=None2023.05 | 42.4 | |
| BERT-baseTraining Method=Fine-tuning, Injection Method=None2023.05 | 41.7 | |
| BERT-baseTraining Method=BitFit, Injection Method=None2023.05 | 41.3 | |
| Entities as Experts2021.10 | 32.5 | |
| KDoSBackbone=Qwen3-4B-Base2026.06 | 16.4 | |
| Diff.Backbone=Qwen3-4B-Base, Synthesis Strategy=Difficulty-weighted2026.06 | 15.2 | |
| Qual.Backbone=Qwen3-4B-Base, Synthesis Strategy=Quality-filtered2026.06 | 14.8 | |
| Uni.Backbone=Qwen3-4B-Base, Synthesis Strategy=Uniform2026.06 | 14.6 | |
| Rand.Backbone=Qwen3-4B-Base, Synthesis Strategy=Random2026.06 | 12.1 | |
| SPBackbone=Qwen3-4B-Base, Training Data=Seed Pool2026.06 | 11.1 | |
| BaseBackbone=Qwen3-4B-Base2026.06 | 10.2 |