List your company

Semantic Textual Similarity on SICK-R (test)

66Similarity ScoreRAIR-Laplacian
37.08844.59452.159.606Mar 24, 2026Apr 6, 2026Apr 20, 2026May 4, 2026May 17, 2026May 31, 2026Jun 14, 2026

Evaluation results

MethodLinks
RAIR-LaplacianHuman comparisons=6002026.0666
Random EloHuman comparisons=6002026.0663.9
SGSHuman comparisons=600, Inferred comparisons=5352026.0663.5
Active EloHuman comparisons=6002026.0662.5
Active Elo (same budget)Human comparisons=6002026.0662.3
Cayley-EncoderBase Model=Llama3-8B, GNN Variant=GIN2026.0360.74
Cayley-EncoderBase Model=Gemma2-2B, GNN Variant=GIN2026.0360.32
FC-EncoderBase Model=Gemma2-2B, GNN Variant=GCN2026.0360.12
Best-LayerBase Model=Pythia-410m2026.0358.26
Cayley-EncoderBase Model=Gemma2-2B, GNN Variant=GCN2026.0357.52
FC-EncoderBase Model=Llama3-8B, GNN Variant=GCN2026.0357.51
Best-LayerBase Model=Gemma2-2B2026.0357.24
FC-EncoderBase Model=Pythia-410m, GNN Variant=GCN2026.0357.14
Best-LayerBase Model=Llama3-8B2026.0356.96
WeightedBase Model=Llama3-8B2026.0356.91
DWAttBase Model=Gemma2-2B2026.0355.36
Cayley-EncoderBase Model=Pythia-410m, GNN Variant=GIN2026.0355.34
WeightedBase Model=Pythia-410m2026.0355.09
FC-EncoderBase Model=Gemma2-2B, GNN Variant=GIN2026.0354.67
Last-LayerBase Model=Pythia-410m2026.0352.55
DWAttBase Model=Pythia-410m2026.0352.43
Cayley-EncoderBase Model=Pythia-410m, GNN Variant=GCN2026.0352.15
FC-EncoderBase Model=Llama3-8B, GNN Variant=GIN2026.0351.85
DWAttBase Model=Llama3-8B2026.0351.45
Last-LayerBase Model=Llama3-8B2026.0351.16
FC-EncoderBase Model=Pythia-410m, GNN Variant=GIN2026.0346.97
WeightedBase Model=Gemma2-2B2026.0346.67
MLP Last-LayerBase Model=Llama3-8B2026.0346.62
Set-EncoderBase Model=Llama3-8B2026.0346.6
Set-EncoderBase Model=Pythia-410m2026.0344.84
Cayley-EncoderBase Model=Llama3-8B, GNN Variant=GCN2026.0344.04
MLP Last-LayerBase Model=Pythia-410m2026.0343.23
Last-LayerBase Model=Gemma2-2B2026.0343.01
Set-EncoderBase Model=Gemma2-2B2026.0342.22
MLP Last-LayerBase Model=Gemma2-2B2026.0340.19
PairS-MSHuman comparisons=6002026.0638.2