Robustness Variability in XLM-R and mBERT Under Sequential Fine-Tuning and StressGAN Adversarial Attacks

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: How does the order of language exposure in sequential fine-tuning affect the robustness of XLM-R and mBERT against StressGAN-generated adversarial examples, measured by F1-score degradation across high-resource (English, Spanish) and low-resource (Yoruba, Turkish) languages?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.6/10.