Universal Dependency Parsing on SpokeBench Discourse+ v1 (test)
71.6ID AccuracyBiLingua Parser
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BiLingua ParserType=LLM-based pipeline2026.02 | 71.6 | 77.4 | 62.6 | 65.6 | 69.8 | |
| DECAPBackbone=GPT-4.12026.02 | 60.4 | 81.3 | 59 | 68.3 | 71 | |
| Bilingual Traditional ParserBackbone=Stanza2026.02 | 29.4 | 29.4 | 29.4 | 29.4 | 39.5 |