Question Answering on MedQA (F1, avg_cost)
72F1 Score (%)Top-k
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Top-kBackbone=Qwen3-8B, Bgt/k=102026.06 | 72 | — | |
| RedundancyBackbone=Qwen3-8B, Bgt/k=302026.06 | 71 | 1.8 | |
| Top-kBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 62 | — | |
| Top-kBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 59 | — | |
| RedundancyBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 59 | 1.68 | |
| RedundancyBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 59 | 1.76 | |
| KnapsackBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 58 | 14.3 | |
| MMRBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 58 | 9.34 | |
| MMRBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 58 | 9.37 | |
| RelevanceBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 57 | 5 | |
| GreedyBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 57 | 13.73 | |
| KnapsackBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 57 | 24.31 | |
| RedundancyBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 57 | 1.76 | |
| Top-kBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 56 | — | |
| Top-kBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 56 | — | |
| RelevanceBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 56 | 0 | |
| RelevanceBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 56 | 14.9 | |
| MMRBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 56 | 7.58 | |
| Vanilla LLMBackbone=Llama3.1-8B-instruct2026.06 | 55 | — | |
| RelevanceBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 55 | 61.7 | |
| GreedyBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 55 | 28.57 | |
| Top-kBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 54 | — | |
| GreedyBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 54 | 18.54 | |
| KnapsackBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 54 | 4.87 | |
| KnapsackBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 54 | 61.73 | |
| RedundancyBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 54 | 1.76 | |
| RelevanceBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 53 | 10 | |
| RelevanceBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 53 | 29.8 | |
| GreedyBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 53 | 0 | |
| KnapsackBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 53 | 9.58 | |
| MMRBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 53 | 8.81 | |
| MMRBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 53 | 9.18 | |
| MMRBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 53 | 9.37 | |
| Top-kBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 52 | — | |
| RelevanceBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 52 | 25 | |
| KnapsackBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 52 | 29.26 | |
| MMRBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 52 | 4.66 | |
| RedundancyBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 52 | 1.18 | |
| GreedyBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 51 | 4.78 | |
| KnapsackBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 51 | 0 | |
| KnapsackBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 51 | 19.2 | |
| RedundancyBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 51 | 1.76 | |
| RedundancyBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 51 | 1.76 | |
| GreedyBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 50 | 23.5 | |
| GreedyBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 50 | 61.73 | |
| GreedyBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 48 | 9.22 | |
| RelevanceBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 47 | 19.9 | |
| MMRBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 47 | 0 | |
| RedundancyBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 47 | 0 | |
| Vanilla LLMBackbone=Qwen3-8B, Bgt/k=-2026.06 | 45 | — |