Semantic Textual Similarity on STS-16
77.63Spearman Rho (x100)BERTlarge-flow (target)
Evaluation Results
| Method | Links | |
|---|---|---|
| BERTlarge-flow (target)Backbone=BERT-large, Pooling Strategy=Last 2 layers average, Flow Training Source=Target dataset2020.11 | 77.63 | |
| BERTbase-flow (target)Backbone=BERT-base, Pooling Strategy=Last 2 layers average, Flow Training Source=Target dataset2020.11 | 75.37 | |
| BERTlarge-flow (NLI*)Backbone=BERT-large, Pooling Strategy=Last 2 layers average, Flow Training Source=NLI corpus2020.11 | 74.47 | |
| MICBackbone=BERT, Dimension=7682026.05 | 72.18 | |
| BERTbase-flow (NLI*)Backbone=BERT-base, Pooling Strategy=Last 2 layers average, Flow Training Source=NLI corpus2020.11 | 71.84 | |
| MIPICBackbone=BERT, Dimension=7682026.04 | 71.56 | |
| BERTlarge-last2avgBackbone=BERT-large, Pooling Strategy=Last 2 layers average2020.11 | 70.32 | |
| BERTbase-last2avgBackbone=BERT-base, Pooling Strategy=Last 2 layers average2020.11 | 69.81 | |
| MICBackbone=TinyBERT 6L, Dimension=7682026.05 | 69.47 | |
| BERTbaseBackbone=BERT-base, Pooling Strategy=Last layer average2020.11 | 65.19 | |
| MICBackbone=BERT, Dimension=162026.05 | 63.76 | |
| Avg. GloVe embeddingsBackbone=GloVe, Pooling Strategy=Average2020.11 | 63.66 | |
| BERTlargeBackbone=BERT-large, Pooling Strategy=Last layer average2020.11 | 61.63 | |
| Avg. BERT embeddingsBackbone=BERT, Pooling Strategy=Average2020.11 | 61.06 | |
| MICBackbone=TinyBERT 6L, Dimension=162026.05 | 60.64 | |
| BERT CLS-vectorBackbone=BERT, Pooling Strategy=[CLS]2020.11 | 38.03 |