@inproceedings{aguda-etal-2025-conservative,
    title = "Conservative Bias in Large Language Models: Measuring Relation Predictions",
    author = "Aguda, Toyin  and
      Wilson, Erik  and
      Anzagira, Allan  and
      Kaur, Simerjot  and
      Smiley, Charese",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/corrections-2025-08/2025.findings-acl.973/",
    doi = "10.18653/v1/2025.findings-acl.973",
    pages = "18989--18998",
    ISBN = "979-8-89176-256-5",
    abstract = "Large language models (LLMs) exhibit pronounced conservative bias in relation extraction tasks, frequently defaulting to no{\_}relation label when an appropriate option is unavailable. While this behavior helps prevent incorrect relation assignments, our analysis reveals that it also leads to significant information loss when reasoning is not explicitly included in the output. We systematically evaluate this trade-off across multiple prompts, datasets, and relation types, introducing the concept of Hobson{'}s choice to capture scenarios where models opt for safe but uninformative labels over hallucinated ones. Our findings suggest that conservative bias occurs twice as often as hallucination. To quantify this effect, we use SBERT and LLM prompts to capture the semantic similarity between conservative bias behaviors in constrained prompts and labels generated from semi-constrained and open-ended prompts."
}