Multiple Choice Question Answering on BT-guided (test)
38.4AccuracyPPT (Distribution)
Evaluation Results
| Method | Links | |
|---|---|---|
| PPT (Distribution)Backbone=QWEN2-7B2026.05 | 38.4 | |
| BaseBackbone=QWEN2-7B2026.05 | 37.6 | |
| PPT (Distribution)Backbone=LLAMA3-8B2026.05 | 37.3 | |
| PPT (Mean)Backbone=LLAMA3-8B2026.05 | 37.2 | |
| PPT (Mean)Backbone=QWEN2-7B2026.05 | 36.9 | |
| BaseBackbone=LLAMA3-8B2026.05 | 35.7 | |
| VerbalizedBackbone=LLAMA3-8B2026.05 | 35.7 |