Lingo_Research_Group at SemEval-2026 Task 9: Evaluating Prompt Variants for Polarization Detection
Researchers from Lingo_Research_Group have detailed their approach for SemEval-2026 Task 9, focusing on multilingual polarization detection. Their study evaluated twelve different prompt designs across three subtasks using the aya-101 and Gemma3-27B models. While effective for coarse-grained polarization detection, the prompt-based methods showed limitations with more nuanced, fine-grained, and multi-label classification tasks. AI
IMPACT Prompt engineering techniques show promise for polarization detection but require further refinement for complex linguistic tasks.