AIxcellent Vibes at GermEval 2025 Shared Task on Candy Speech Detection: Improving Model Performance by Span-Level Training
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914039548870656 |
|---|---|
| author | Thelen, Christian Rene Blaneck, Patrick Gustav Bornheim, Tobias Grieger, Niklas Bialonski, Stephan |
| author_facet | Thelen, Christian Rene Blaneck, Patrick Gustav Bornheim, Tobias Grieger, Niklas Bialonski, Stephan |
| contents | Positive, supportive online communication in social media (candy speech) has the potential to foster civility, yet automated detection of such language remains underexplored, limiting systematic analysis of its impact. We investigate how candy speech can be reliably detected in a 46k-comment German YouTube corpus by monolingual and multilingual language models, including GBERT, Qwen3 Embedding, and XLM-RoBERTa. We find that a multilingual XLM-RoBERTa-Large model trained to detect candy speech at the span level outperforms other approaches, ranking first in both binary positive F1: 0.8906) and categorized span-based detection (strict F1: 0.6307) subtasks at the GermEval 2025 Shared Task on Candy Speech Detection. We speculate that span-based training, multilingual capabilities, and emoji-aware tokenizers improved detection performance. Our results demonstrate the effectiveness of multilingual models in identifying positive, supportive language. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_07459 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AIxcellent Vibes at GermEval 2025 Shared Task on Candy Speech Detection: Improving Model Performance by Span-Level Training Thelen, Christian Rene Blaneck, Patrick Gustav Bornheim, Tobias Grieger, Niklas Bialonski, Stephan Computation and Language Positive, supportive online communication in social media (candy speech) has the potential to foster civility, yet automated detection of such language remains underexplored, limiting systematic analysis of its impact. We investigate how candy speech can be reliably detected in a 46k-comment German YouTube corpus by monolingual and multilingual language models, including GBERT, Qwen3 Embedding, and XLM-RoBERTa. We find that a multilingual XLM-RoBERTa-Large model trained to detect candy speech at the span level outperforms other approaches, ranking first in both binary positive F1: 0.8906) and categorized span-based detection (strict F1: 0.6307) subtasks at the GermEval 2025 Shared Task on Candy Speech Detection. We speculate that span-based training, multilingual capabilities, and emoji-aware tokenizers improved detection performance. Our results demonstrate the effectiveness of multilingual models in identifying positive, supportive language. |
| title | AIxcellent Vibes at GermEval 2025 Shared Task on Candy Speech Detection: Improving Model Performance by Span-Level Training |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2509.07459 |