Semantic Textual Similarity on STS16 (test)
77.18Spearman CorrDINO (STS-A-x2)
Evaluation Results
| Method | Links | |
|---|---|---|
| DINO (STS-A-x2)supervision=unsupervised, unlabeled_data=STS2021.04 | 77.18 | |
| DINO (STS-A-x1x2)supervision=unsupervised2021.04 | 76.47 | |
| Li et al. (2020)supervision=unsupervised, unlabeled_data=STS2021.04 | 75.37 | |
| S-BERT (base)supervision=supervised2021.04 | 74.3 | |
| S-ROBERTa (base)supervision=supervised2021.04 | 73.69 | |
| USEsupervision=supervised2021.04 | 73.18 | |
| Li et al. (2020)supervision=unsupervised, unlabeled_data=NLI2021.04 | 71.84 | |
| Zhang et al. (2020)supervision=unsupervised, unlabeled_data=NLI2021.04 | 70.16 | |
| InferSent, Glovesupervision=supervised2021.04 | 66.87 | |
| Avg. GloVesupervision=unsupervised2021.04 | 63.66 | |
| Cayley-EncoderBase Model=Llama3-8B, GNN Variant=GIN2026.03 | 63.13 | |
| FC-EncoderBase Model=Gemma2-2B, GNN Variant=GCN2026.03 | 62.96 | |
| Cayley-EncoderBase Model=Gemma2-2B, GNN Variant=GIN2026.03 | 62.69 | |
| Avg. BERTsupervision=unsupervised2021.04 | 61.06 | |
| Best-LayerBase Model=Llama3-8B2026.03 | 59.04 | |
| WeightedBase Model=Llama3-8B2026.03 | 58.97 | |
| Best-LayerBase Model=Gemma2-2B2026.03 | 58.89 | |
| Cayley-EncoderBase Model=Pythia-410m, GNN Variant=GIN2026.03 | 58.63 | |
| Best-LayerBase Model=Pythia-410m2026.03 | 58.02 | |
| FC-EncoderBase Model=Gemma2-2B, GNN Variant=GIN2026.03 | 57.28 | |
| FC-EncoderBase Model=Llama3-8B, GNN Variant=GCN2026.03 | 57.25 | |
| Cayley-EncoderBase Model=Gemma2-2B, GNN Variant=GCN2026.03 | 57.24 | |
| FC-EncoderBase Model=Pythia-410m, GNN Variant=GCN2026.03 | 57.11 | |
| Cayley-EncoderBase Model=Pythia-410m, GNN Variant=GCN2026.03 | 56.62 | |
| WeightedBase Model=Gemma2-2B2026.03 | 54.91 | |
| Last-LayerBase Model=Llama3-8B2026.03 | 54.32 | |
| FC-EncoderBase Model=Llama3-8B, GNN Variant=GIN2026.03 | 53.89 | |
| WeightedBase Model=Pythia-410m2026.03 | 52.76 | |
| MLP Last-LayerBase Model=Llama3-8B2026.03 | 52.55 | |
| Last-LayerBase Model=Gemma2-2B2026.03 | 51.4 | |
| MLP Last-LayerBase Model=Gemma2-2B2026.03 | 51.16 | |
| DWAttBase Model=Gemma2-2B2026.03 | 51.11 | |
| Last-LayerBase Model=Pythia-410m2026.03 | 50.37 | |
| FC-EncoderBase Model=Pythia-410m, GNN Variant=GIN2026.03 | 49.56 | |
| DWAttBase Model=Llama3-8B2026.03 | 48.86 | |
| Set-EncoderBase Model=Gemma2-2B2026.03 | 47.34 | |
| DWAttBase Model=Pythia-410m2026.03 | 44.8 | |
| Cayley-EncoderBase Model=Llama3-8B, GNN Variant=GCN2026.03 | 42.09 | |
| Set-EncoderBase Model=Pythia-410m2026.03 | 40.48 | |
| MLP Last-LayerBase Model=Pythia-410m2026.03 | 39.84 | |
| Set-EncoderBase Model=Llama3-8B2026.03 | 39.5 | |
| BERT CLSsupervision=unsupervised2021.04 | 38.08 |