Generative Question Answering on WebQuestions
52HALL ScoreBase
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| BaseBackbone=Llama 3.1-8b2026.05 | 52 | 3 | 45 | 45.69 | |
| Append-idkBackbone=Llama 3.1-8b2026.05 | 48 | 6.33 | 45.67 | 47.16 | |
| EmoPrompt (avg)Backbone=Llama 3.1-8b2026.05 | 46.03 | 7.82 | 46.15 | 48 | |
| Plan-and-SolveBackbone=Gemma-3-12b-pt2026.05 | 45.33 | 2.67 | 52 | 52.7 | |
| RCSPBackbone=Llama 3.1-8b2026.05 | 41.33 | 7 | 51.67 | 53.5 | |
| Plan-and-SolveBackbone=Llama 3.1-8b2026.05 | 40.67 | 9 | 50.33 | 52.71 | |
| BaseBackbone=Gemma-3-12b-pt2026.05 | 40 | 5 | 55 | 56.41 | |
| Append-idkBackbone=Gemma-3-12b-pt2026.05 | 39.67 | 6.33 | 54 | 55.77 | |
| EmoPrompt (avg)Backbone=Gemma-3-12b-pt2026.05 | 39.58 | 7.91 | 52.51 | 54.68 | |
| RCSPBackbone=Gemma-3-12b-pt2026.05 | 37.33 | 6.67 | 56 | 57.93 |