Definition Modeling on Wiki (BLEU, ROUGE)
62.07BLEULM-LEXICON-DENSE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LM-LEXICON-DENSEParameter Count=8B, Mode=BoN-Oracle2026.02 | 62.07 | 68.62 | |
| LM-LEXICON-MOEParameter Count=4x8B, Mode=BoN-ORM2026.02 | 60.88 | 57.66 | |
| LM-LEXICON-MOEParameter Count=4x8B, Mode=Zero-shot2026.02 | 60.31 | 55.52 | |
| LM-LEXICON-DENSEParameter Count=8B, Mode=BoN-ORM2026.02 | 59.33 | 59.46 | |
| LM-LEXICON-DENSEParameter Count=8B, Mode=Zero-shot2026.02 | 57.92 | 59.56 | |
| Rerank-T5Supervised training=true2026.02 | 55.61 | 57.25 | |
| Contrast-T5Supervised training=true2026.02 | 55.26 | 42.27 | |
| MDM-T5Supervised training=true2026.02 | 54.33 | 55.83 | |
| LM-LEXICON-MOEParameter Count=4x8B, Mode=BoN-Oracle2026.02 | 51.62 | 61.88 | |
| SimpDefinerSupervised training=true2026.02 | 44.03 | 49.26 | |
| GEMINI-1.5-PROIn-Context Learning=Random-ICL, shots=322026.02 | 36.87 | 48.13 | |
| GEMINI-1.5-PROIn-Context Learning=Retrieval-ICL, shots=322026.02 | 35.59 | 43.71 | |
| GPT-4-TURBOIn-Context Learning=Retrieval-ICL, shots=322026.02 | 35.4 | 40.68 | |
| GPT-4-TURBOIn-Context Learning=Random-ICL, shots=322026.02 | 31.63 | 45.89 | |
| CLAUDE-3-OPUSIn-Context Learning=Random-ICL, shots=322026.02 | 23.3 | 35.22 | |
| CLAUDE-3-OPUSIn-Context Learning=Retrieval-ICL, shots=322026.02 | 14.59 | 15.83 | |
| LlamaDictSupervised training=true2026.02 | 14.38 | 25.29 | |
| Flan-T5-DefSupervised training=true2026.02 | 13.82 | 23.97 |