Word Sense Induction on SemEval 2010
43.6V-MPolyLM large
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| PolyLM largeSize=Large2026.03 | 43.6 | 67.5 | 6.2 | 49.2 | |
| PolyLM baseSize=Base2026.03 | 41.8 | 66.4 | 6.2 | 49.1 | |
| LSDPSubstitutes=BERT-large2026.03 | 38.9 | 70.7 | 4.6 | 52.8 | |
| GPT-4oMethodology=LLM-based WSI prompting2026.03 | 36.3 | 63.9 | 7.1 | 47.7 | |
| 1cpexStrategy=One cluster per example2026.03 | 31.7 | 0 | 19.5 | 8 | |
| Llama 3.3 70BMethodology=LLM-based WSI prompting, Parameters=70B2026.03 | 29.4 | 49.7 | 8.1 | 49.6 | |
| Llama 3.1 8BMethodology=LLM-based WSI prompting, Parameters=8B2026.03 | 16.5 | 49.3 | 7.3 | 49.6 | |
| 1cplStrategy=One cluster per lemma2026.03 | 0 | 63.5 | 0 | 64.1 |