DeBERTa-v3-large: Joint Sentence + LLM Definition Training
Cross-encoder for Word-in-Context (WiC) that jointly encodes marked sentences and LLM-generated sense definitions.
Input Format
text: word in sentence1 [SEP] definition1 text_pair: word in sentence2 [SEP] definition2
Performance
| Split | Accuracy |
|---|---|
| Validation | 0.7555 |
| Test | 0.7179 |
| SenseBERT (SOTA) | 0.7210 |
- Downloads last month
- 48
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support