List your company

Multilingual Question Answering on TyDiQA (1-shot macro-averaged)

48.86F1 Score (1-shot macro)Filter-then-Weight
11.243221.009130.77540.5409Mar 8, 2026

Evaluation results

MethodLinks
Filter-then-WeightBackbone=Llama-3.2-1B, Selection Protocol=Best-of-run2026.0348.86
Filter-then-WeightBackbone=Llama-3.2-1B, Selection Protocol=Fixed-budget2026.0348.67
GREATSBackbone=Llama-3.2-1B, Selection Protocol=Best-of-run2026.0347.5
LESSBackbone=Llama-3.2-1B, Selection Protocol=Best-of-run2026.0347.21
GREATSBackbone=Llama-3.2-1B, Selection Protocol=Fixed-budget2026.0347.16
TracIncBackbone=Llama-3.2-1B, Selection Protocol=Best-of-run2026.0347.01
TracIncBackbone=Llama-3.2-1B, Selection Protocol=Fixed-budget2026.0347.01
LESSBackbone=Llama-3.2-1B, Selection Protocol=Fixed-budget2026.0346.88
Filter-then-WeightBackbone=Qwen3-0.6B, Selection Protocol=Best-of-run2026.0346.29
GRAD-MATCHBackbone=Llama-3.2-1B, Selection Protocol=Best-of-run2026.0345.35
GREATSBackbone=Qwen3-0.6B, Selection Protocol=Best-of-run2026.0345.25
TracIncBackbone=Qwen3-0.6B, Selection Protocol=Best-of-run2026.0345.17
LESSBackbone=Qwen3-0.6B, Selection Protocol=Best-of-run2026.0345.07
Filter-then-WeightBackbone=Qwen3-0.6B, Selection Protocol=Fixed-budget2026.0344.82
GRAD-MATCHBackbone=Qwen3-0.6B, Selection Protocol=Best-of-run2026.0344.23
GRAD-MATCHBackbone=Llama-3.2-1B, Selection Protocol=Fixed-budget2026.0344.15
FullBackbone=Qwen3-0.6B, Selection Protocol=Best-of-run2026.0343.65
LESSBackbone=Qwen3-0.6B, Selection Protocol=Fixed-budget2026.0343.57
GRAD-MATCHBackbone=Qwen3-0.6B, Selection Protocol=Fixed-budget2026.0343.14
TracIncBackbone=Qwen3-0.6B, Selection Protocol=Fixed-budget2026.0343.07
GREATSBackbone=Qwen3-0.6B, Selection Protocol=Fixed-budget2026.0342.7
FullBackbone=Llama-3.2-1B, Selection Protocol=Best-of-run2026.0342.47
FullBackbone=Qwen3-0.6B, Selection Protocol=Fixed-budget2026.0341.14
FullBackbone=Llama-3.2-1B, Selection Protocol=Fixed-budget2026.0338.68
OriginBackbone=Qwen3-0.6B, Selection Protocol=Best-of-run2026.0320.94
OriginBackbone=Qwen3-0.6B, Selection Protocol=Fixed-budget2026.0320.94
OriginBackbone=Llama-3.2-1B, Selection Protocol=Best-of-run2026.0312.69
OriginBackbone=Llama-3.2-1B, Selection Protocol=Fixed-budget2026.0312.69