Universal Dependency Parsing on SpokeBench Overall v1 (test)
76.2ID AccuracyBiLingua Parser
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BiLingua ParserType=LLM-based pipeline2026.02 | 76.2 | 78.5 | 62.3 | 65.1 | 72.2 | |
| DECAPBackbone=GPT-4.12026.02 | 58.2 | 73.7 | 54.6 | 59.1 | 66.6 | |
| Bilingual Traditional ParserBackbone=Stanza2026.02 | 15.6 | 15.7 | 15.7 | 15.7 | 29.5 |