Universal Dependency Parsing on SpokeBench Repetition v1 (test)
78ID ScoreBiLingua Parser
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BiLingua ParserType=LLM-based pipeline2026.02 | 78 | 85 | 65 | 63.5 | 74.7 | |
| DECAPBackbone=GPT-4.1 (version 2025-04-14), Decoding=deterministic (temperature = 0)2026.02 | 72 | 85.3 | 68 | 71 | 77.7 | |
| Bilingual Traditional ParserBackbone=Stanza, Training data=Mixture of English and Spanish2026.02 | 9.7 | 9.7 | 9.7 | 9.7 | 24.8 |