Sentiment Classification on Stanford Sentiment Treebank SST-5 (test)
54.7AccuracyBCN + Char + ELMo
Evaluation Results
| Method | Links | |
|---|---|---|
| BCN + Char + ELMopre-trained on large external corpora=true2018.09 | 54.7 | |
| Gumbel Tree-LSTM2018.09 | 53.7 | |
| Bi-CAS-LSTMlayers=22018.09 | 53.6 | |
| CAS-LSTMlayers=22018.09 | 53 | |
| byte mLSTMpre-trained on large external corpora=true2018.09 | 52.9 | |
| Constituency Tree-LSTMrecurrent dropout=true2018.09 | 52.3 | |
| Constituency Tree-LSTM2018.09 | 51 | |
| BiLSTMlayers=22018.09 | 48.5 | |
| LSTMlayers=22018.09 | 46 | |
| Recursive Neural Tensor Network2018.09 | 45.7 | |
| Switch-ACMoEPre-training dataset=WikiText-103, Task=Finetuning2025.02 | 40.04 | |
| Switch TransformerPre-training dataset=WikiText-103, Task=Finetuning2025.02 | 39.13 | |
| GLaM-ACMoEPre-training dataset=WikiText-103, Task=Finetuning2025.02 | 34.24 | |
| Switch-ACMoEPre-training dataset=EnWik-8, Task=Finetuning2025.02 | 33.79 | |
| GLaMPre-training dataset=WikiText-103, Task=Finetuning2025.02 | 33.69 | |
| Switch TransformerPre-training dataset=EnWik-8, Task=Finetuning2025.02 | 32.21 |