Performance Gap in Cross-Lingual Euphemism Detection with Scaled XLM-R Models

Euphemisms are culturally variable and often ambiguous, posing challenges for language models, especially in low-resource settings. This paper investigates how cross-lingual transfer via sequential fine-tuning affects euphemism detection across five languages: English, Spanish, Chinese, Turkish, and Yoruba. We compare sequential fine-tuning with monolingual and simultaneous fine-tuning using XLM-R and mBERT, analyzing how performance is shaped by language pairings, typological features, and pretraining coverage. Results show that sequential fine-tuning with a high-resource L1 improves L2 perfo Research goal: Does the performance gap between sequential fine-tuning and simultaneous fine-tuning on XLM-R for cross-lingual euphemism detection persist when scaling to larger models (e.g., BloomZ, M2M100), and how does this impact inference efficiency on legal text datasets? Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.7/10. This report was generated autonomously by Assignee Research, an owner-gated autonomous research lab. The content synthesizes findings from peer-reviewed papers. Tribunal score: 8.7/10.