大型语言模型中的保守偏见：衡量关系预测
Conservative Bias in Large Language Models: Measuring Relation Predictions
AI总结：
本文研究大型语言模型在关系抽取中偏好 No_Relation 的保守偏见，通过多提示、数据集和关系类型评估，并用 SBERT 与 LLM 提示量化其相对幻觉更频繁的语义表现。
摘要：
大型语言模型（LLM）在关系抽取任务中表现出明显的保守偏见，当没有合适选项时，往往默认选择 No_Relation 标签。虽然这种行为有助于防止错误的关系分配，但我们的分析表明，当输出中未明确包含推理时，它也会导致严重的信息损失。我们在多种提示、数据集和关系类型上系统评估了这一权衡，并引入 Hobson’s choice（霍布森选择）概念，用以描述模型选择安全但无信息量的标签，而非幻觉标签的情形。研究结果表明，保守偏见的发生频率是幻觉的两倍。为量化该效应，我们使用 SBERT 和 LLM 提示来捕捉受限提示中的保守偏见行为，与半受限和开放式提示生成的标签之间的语义相似性。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
