List your company

Question Answering on MMLU (F1 and Cost Evaluation)

84F1 ScoreMMR
53.8461.6769.577.33Jun 1, 2026

Evaluation results

MethodLinks
MMRBackbone=Qwen3-8B, Bgt/k=252026.06849.6
KnapsackBackbone=Qwen3-8B, Bgt/k=52026.06834.9
KnapsackBgt/k=30, Backbone=Llama3.1-8B-instruct2026.067927.68
GreedyBgt/k=25, Backbone=Llama3.1-8B-instruct2026.067622.99
GreedyBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.067548.24
Vanilla LLMBackbone=Qwen3-8B, Bgt/k=-2026.0675—
Top-kBgt/k=15, Backbone=Llama3.1-8B-instruct2026.0672—
KnapsackBgt/k=20, Backbone=Llama3.1-8B-instruct2026.067218.97
KnapsackBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.067248.24
Top-kBgt/k=30, Backbone=Llama3.1-8B-instruct2026.0671—
RelevanceBgt/k=15, Backbone=Llama3.1-8B-instruct2026.067114.7
GreedyBgt/k=30, Backbone=Llama3.1-8B-instruct2026.067127.08
RelevanceBgt/k=25, Backbone=Llama3.1-8B-instruct2026.067024.1
GreedyBgt/k=20, Backbone=Llama3.1-8B-instruct2026.067018.16
MMRBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06704.62
KnapsackBgt/k=15, Backbone=Llama3.1-8B-instruct2026.066914.1
KnapsackBgt/k=25, Backbone=Llama3.1-8B-instruct2026.066923.59
MMRBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06697.85
Top-kBgt/k=20, Backbone=Llama3.1-8B-instruct2026.0668—
RedundancyBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06680
Vanilla LLMBackbone=Llama3.1-8B-instruct2026.0667—
Top-kBgt/k=25, Backbone=Llama3.1-8B-instruct2026.0667—
GreedyBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06670
GreedyBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06679.13
GreedyBgt/k=15, Backbone=Llama3.1-8B-instruct2026.066713.68
RelevanceBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06669.9
RelevanceBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.066648.2
MMRBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06669.12
MMRBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06669.63
MMRBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06669.68
RedundancyBgt/k=15, Backbone=Llama3.1-8B-instruct2026.06660.91
RelevanceBgt/k=30, Backbone=Llama3.1-8B-instruct2026.066528.1
MMRBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06659.68
RedundancyBgt/k=25, Backbone=Llama3.1-8B-instruct2026.06650.91
Top-kBgt/k=5, Backbone=Llama3.1-8B-instruct2026.0664—
KnapsackBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06644.88
RedundancyBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06640.91
RedundancyBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06640.91
RedundancyBgt/k=∞, Backbone=Llama3.1-8B-instruct2026.06641.76
Top-kBgt/k=10, Backbone=Llama3.1-8B-instruct2026.0663—
MMRBgt/k=20, Backbone=Llama3.1-8B-instruct2026.06639.52
RedundancyBgt/k=30, Backbone=Llama3.1-8B-instruct2026.06630.91
RelevanceBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06625
RelevanceBgt/k=20, Backbone=Llama3.1-8B-instruct2026.066219.6
RelevanceBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06610
KnapsackBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06600
RedundancyBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06590.64
GreedyBgt/k=5, Backbone=Llama3.1-8B-instruct2026.06584.82
KnapsackBgt/k=10, Backbone=Llama3.1-8B-instruct2026.06579.55
MMRBgt/k=0, Backbone=Llama3.1-8B-instruct2026.06550