@inproceedings{rafsan-2026-sentiment,
    title = "Sentiment Syndicate at {S}em{E}val-2026 Task 6: Reframing Political Question{--}Answer Interactions via Natural Language Inference for Clarity Level Classification",
    author = "Rafsan, Rafi",
    editor = "Kochmar, Ekaterina  and
      Ghosh, Debanjan  and
      North, Kai  and
      Komachi, Mamoru",
    booktitle = "Proceedings of the 20th {I}nternational {W}orkshop on {S}emantic {E}valuation (2026)",
    month = jul,
    year = "2026",
    address = "San Diego, California, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.semeval-1.350/",
    pages = "2779--2786",
    ISBN = "979-8-89176-414-9",
    abstract = "This paper presents the Sentiment Syndicate team{'}s submission to SemEval-2026 Task 6, Subtask 1 (CLARITY: Unmasking Political Question Evasions), which focuses on classifying the clarity level of political question{--}answer interactions. We investigate three modeling strategies: (1) fine-tuning a RoBERTa-based classifier, (2) reformulating the task as a Natural Language Inference (NLI) problem, and (3) leveraging large language models (LLMs) for classification. All approaches are evaluated using macro F1-score on the official dataset. Experimental results demonstrate that the NLI based formulation outperforms the other strategies, highlighting the effectiveness of modeling semantic alignment between questions and answers. Our best-performing system achieves an F1-score of 0.67 on the test set."
}