Generative Question Answering on HaluEval (test)
50.33HALL RatePlan-and-Solve
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Plan-and-SolveBackbone=Llama 3.1-8b2026.05 | 50.33 | 17.67 | 32 | 35.1 | |
| Append-idkBackbone=Gemma-3-12b-pt2026.05 | 54.67 | 9.33 | 36 | 37.76 | |
| RCSPBackbone=Gemma-3-12b-pt2026.05 | 55 | 6 | 39 | 40.21 | |
| EmoPrompt (avg)Backbone=Gemma-3-12b-pt2026.05 | 55.79 | 8.73 | 35.48 | 37.1 | |
| Plan-and-SolveBackbone=Gemma-3-12b-pt2026.05 | 56.67 | 5.67 | 37.67 | 38.77 | |
| BaseBackbone=Gemma-3-12b-pt2026.05 | 58 | 4 | 38 | 38.77 | |
| BaseBackbone=Llama 3.1-8b2026.05 | 59.67 | 6.33 | 34 | 35.11 | |
| Append-idkBackbone=Llama 3.1-8b2026.05 | 60 | 4.33 | 35.67 | 36.46 | |
| RCSPBackbone=Llama 3.1-8b2026.05 | 60.67 | 1.33 | 38 | 38.26 | |
| EmoPrompt (avg)Backbone=Llama 3.1-8b2026.05 | 63.15 | 6.67 | 30.18 | 31.25 |