Semantic Textual Similarity on STS12 (test)
65.31Similarity ScoreCayley-Encoder
Evaluation Results
| Method | Links | |
|---|---|---|
| Cayley-EncoderBase Model=Llama3-8B, GNN Variant=GIN2026.03 | 65.31 | |
| Cayley-EncoderBase Model=Gemma2-2B, GNN Variant=GCN2026.03 | 64.36 | |
| FC-EncoderBase Model=Gemma2-2B, GNN Variant=GCN2026.03 | 61.83 | |
| Cayley-EncoderBase Model=Gemma2-2B, GNN Variant=GIN2026.03 | 60.38 | |
| Cayley-EncoderBase Model=Pythia-410m, GNN Variant=GCN2026.03 | 57.9 | |
| Cayley-EncoderBase Model=Pythia-410m, GNN Variant=GIN2026.03 | 55.97 | |
| FC-EncoderBase Model=Llama3-8B, GNN Variant=GCN2026.03 | 55.19 | |
| DWAttBase Model=Pythia-410m2026.03 | 52.51 | |
| Set-EncoderBase Model=Pythia-410m2026.03 | 52.11 | |
| DWAttBase Model=Gemma2-2B2026.03 | 51.86 | |
| DWAttBase Model=Llama3-8B2026.03 | 51.47 | |
| Best-LayerBase Model=Pythia-410m2026.03 | 50.62 | |
| FC-EncoderBase Model=Pythia-410m, GNN Variant=GCN2026.03 | 50.13 | |
| Best-LayerBase Model=Llama3-8B2026.03 | 47.21 | |
| WeightedBase Model=Pythia-410m2026.03 | 47.12 | |
| Cayley-EncoderBase Model=Llama3-8B, GNN Variant=GCN2026.03 | 47.11 | |
| Last-LayerBase Model=Pythia-410m2026.03 | 46.96 | |
| FC-EncoderBase Model=Gemma2-2B, GNN Variant=GIN2026.03 | 46.48 | |
| FC-EncoderBase Model=Pythia-410m, GNN Variant=GIN2026.03 | 44.69 | |
| Set-EncoderBase Model=Gemma2-2B2026.03 | 43.97 | |
| Best-LayerBase Model=Gemma2-2B2026.03 | 43.02 | |
| MLP Last-LayerBase Model=Pythia-410m2026.03 | 42.14 | |
| FC-EncoderBase Model=Llama3-8B, GNN Variant=GIN2026.03 | 41.39 | |
| Set-EncoderBase Model=Llama3-8B2026.03 | 39.94 | |
| WeightedBase Model=Gemma2-2B2026.03 | 35.66 | |
| MLP Last-LayerBase Model=Llama3-8B2026.03 | 34.88 | |
| Last-LayerBase Model=Gemma2-2B2026.03 | 33.7 | |
| WeightedBase Model=Llama3-8B2026.03 | 32.86 | |
| Last-LayerBase Model=Llama3-8B2026.03 | 32.65 | |
| MLP Last-LayerBase Model=Gemma2-2B2026.03 | 27.36 |