Few-shot text-based emotion detection
This work addresses emotion detection for text in low-resource languages, but it is incremental as it applies existing methods to new data in a competition setting.
The paper tackled few-shot text-based emotion detection in multiple languages by experimenting with large language models like Gemini, Qwen, and DeepSeek using few-shot prompting or fine-tuning, achieving an F1-macro of 0.7546 in English (26th out of 96 teams), 0.1727 in Portuguese (35th out of 36 teams), and 0.325 in Emakhuwa (1st out of 31 teams).
This paper describes the approach of the Unibuc - NLP team in tackling the SemEval 2025 Workshop, Task 11: Bridging the Gap in Text-Based Emotion Detection. We mainly focused on experiments using large language models (Gemini, Qwen, DeepSeek) with either few-shot prompting or fine-tuning. With our final system, for the multi-label emotion detection track (track A), we got an F1-macro of $0.7546$ (26/96 teams) for the English subset, $0.1727$ (35/36 teams) for the Portuguese (Mozambican) subset and $0.325$ (\textbf{1}/31 teams) for the Emakhuwa subset.