Multi-word Expression Identification on CoAM (test)
70PrecisionDLT+la
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DLT+laBackbone=DeBERTa-v3-large, Linguistic Feature Integration=Enabled, Data Augmentation=Lexical Substitution2026.01 | 70 | 66.4 | 68.2 | |
| DLT+loBackbone=DeBERTa-v3-large, Linguistic Feature Integration=Enabled, Data Augmentation=Oversampling2026.01 | 69.3 | 70.3 | 69.8 | |
| DBT+aBackbone=DeBERTa-v3-large, Data Augmentation=Lexical Substitution2026.01 | 68.2 | 58.5 | 63 | |
| BLT2026.01 | 67.4 | 53.3 | 59.5 | |
| DBT+laBackbone=DeBERTa-v3-large, Linguistic Feature Integration=Enabled, Data Augmentation=Lexical Substitution2026.01 | 66.4 | 63.8 | 65.1 | |
| DBT+oBackbone=DeBERTa-v3-large, Data Augmentation=Oversampling2026.01 | 65 | 61.9 | 63.4 | |
| DLTBackbone=DeBERTa-v3-large2026.01 | 64 | 61.2 | 62.6 | |
| DBT+lBackbone=DeBERTa-v3-large, Linguistic Feature Integration=Enabled2026.01 | 64 | 68 | 65.9 | |
| Qwen-72BModel Size=72B2026.01 | 63.8 | 52.8 | 57.8 | |
| DBT+loBackbone=DeBERTa-v3-large, Linguistic Feature Integration=Enabled, Data Augmentation=Oversampling2026.01 | 63.6 | 66.4 | 65 | |
| BBT2026.01 | 62.5 | 55.6 | 58.9 | |
| DBTBackbone=DeBERTa-v3-large2026.01 | 62.1 | 65.4 | 63.7 | |
| DLT+oBackbone=DeBERTa-v3-large, Data Augmentation=Oversampling2026.01 | 61.3 | 63.5 | 62.4 | |
| DLT+aBackbone=DeBERTa-v3-large, Data Augmentation=Lexical Substitution2026.01 | 61.1 | 69.3 | 64.9 | |
| DLT+lBackbone=DeBERTa-v3-large, Linguistic Feature Integration=Enabled2026.01 | 52.6 | 69.3 | 59.8 | |
| BBS2026.01 | 51.6 | 55.9 | 53.7 |