Question Answering on TriviaQA (F1, avg_cost)
64.47F1 ScoreRelevance
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RelevanceBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 64.47 | 18 | |
| KnapsackBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 64 | 12.3 | |
| MMRBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 63.78 | 9.87 | |
| GreedyBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 63.71 | 19.38 | |
| MMRBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 63.55 | 11.18 | |
| Top-kBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 63.23 | — | |
| MMRBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 62.89 | 8 | |
| KnapsackBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 62.57 | 22.3 | |
| KnapsackBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 62.29 | 19.59 | |
| GreedyBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 62.12 | 14.94 | |
| MMRBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 61.92 | 10.86 | |
| MMRBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 61.83 | 11.18 | |
| RelevanceBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 61.81 | 22.3 | |
| GreedyBackbone=Qwen3-8B, Bgt/k=302026.06 | 61.79 | 19.4 | |
| KnapsackBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 61.69 | 17.81 | |
| GreedyBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 61.44 | 12.03 | |
| RelevanceBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 61.19 | 0 | |
| GreedyBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 61.18 | 17.6 | |
| MMRBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 61.07 | 11.16 | |
| RelevanceBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 60.93 | 12.6 | |
| RelevanceBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 60.87 | 4.9 | |
| RelevanceBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 60.62 | 19.7 | |
| Top-kBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 60.4 | — | |
| MMRBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 60.37 | 4.73 | |
| GreedyBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 60.34 | 22.3 | |
| RelevanceBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 60.29 | 15.6 | |
| KnapsackBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 60.05 | 15.45 | |
| Top-kBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 59.45 | — | |
| GreedyBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 59.36 | 8.78 | |
| RelevanceBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 59.18 | 9.1 | |
| Top-kBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 59.13 | — | |
| RedundancyBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06 | 59.05 | 0.09 | |
| KnapsackBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 58.87 | 4.82 | |
| RedundancyBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 58.66 | 0 | |
| KnapsackBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 58.5 | 8.87 | |
| RedundancyBackbone=Qwen3-8B, Bgt/k=202026.06 | 58.46 | 0.1 | |
| RedundancyBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06 | 58.24 | 0.09 | |
| GreedyBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 58.03 | 0 | |
| RedundancyBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06 | 57.66 | 0.09 | |
| RedundancyBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06 | 57.6 | 0.09 | |
| RedundancyBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06 | 57.42 | 0.09 | |
| RedundancyBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 57.41 | 0.09 | |
| MMRBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 56.99 | 0 | |
| GreedyBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 56.84 | 4.82 | |
| RedundancyBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 56.26 | 0.09 | |
| KnapsackBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06 | 56.13 | 0 | |
| Top-kBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06 | 54.38 | — | |
| Top-kBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06 | 47.41 | — | |
| Vanilla LLMBackbone=Qwen3-8B, Bgt/k=-2026.06 | 43.57 | — | |
| Vanilla LLMBackbone=Llama3.1-8B-instruct2026.06 | 24.28 | — |