发表机构
Hong Kong Baptist University（香港浸会大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对LLM忠实性评估的不足，提出成对上下文敏感基准PFaithBench，发现模型存在回答偏向，训练数据组成影响忠实性权衡。
AI 中文摘要
大型语言模型（LLMs）被期望基于提供的上下文忠实地回答问题，当上下文信息不足以回答问题时，应弃权（不执行）。现有的忠实性评估通常孤立地评估每个问题-上下文实例；然而，这种实例级评估未能捕捉忠实行为的一个基本要求：根据可用上下文的变化调整模型响应的能力。特别是，当存在充分证据时，模型应提供正确答案，而当证据不足时应弃权（不执行）。在这项工作中，我们提出了一个成对忠实性基准（PFaithBench），用于评估模型在支持性上下文与非支持性上下文下，对于同一问题能否在回答与弃权（不执行）之间切换。我们对七个模型家族的三十九个模型的评估表明，忠实性本质上涉及回答与弃权（不执行）之间的权衡，且大多数当前模型表现出强烈的回答偏向，大多数忠实性错误源于过度回答，即模型即使在提供的上下文不足时也倾向于编造响应。我们进一步在不同数据构建下进行了一系列关于忠实性训练的研究。我们的结果表明，训练结果对回答和弃权（不执行）数据的具体组成高度敏感。从不匹配的来源构建回答和弃权（不执行）数据可能导致模型依赖数据集特定的捷径而非实际的上下文充分性。此外，增加回答监督数据提高了回答性能但加剧了过度回答，而增加弃权（不执行）数据减少了幻觉但导致了过度弃权（不执行）。代码和数据已在此https URL发布。
英文摘要
Large language models (LLMs) are expected to answer questions faithfully based on the provided context, abstaining when the context information is insufficient to answer the questions. Existing faithfulness evaluations typically assess each question-context instance in isolation; however, such instance-level evaluation fails to capture a fundamental requirement of faithful behavior: the ability to adapt model responses to changes in available contexts. In particular, a model should provide correct answers when sufficient evidence is present and abstain when it is not. In this work, we propose a Pairwise Faithfulness Benchmark (PFaithBench) that evaluates whether a model can switch between answering and abstaining for the same question under supporting versus non-supporting contexts. Our evaluations across thirty-nine models with seven model families demonstrate that faithfulness fundamentally involves a trade-off between answering and abstaining, and that most current models exhibit a strong bias toward answering, with most faithfulness errors arising from over-answering, i.e., models tend to fabricate a response even when the provided context is insufficient. We further conduct a series of studies on faithfulness training under different data constructions. Our results show that training outcomes are highly sensitive to the specific composition of answering and abstaining data. Constructing answering and abstaining data from mismatched sources can cause models to rely on dataset-specific shortcuts rather than actual context sufficiency. Moreover, increasing answer-supervised data improves answering performance but exacerbates over-answering, while increasing abstaining data reduces hallucination but leads to over-abstention. The code and data are released at https://github.com/tmlr-group/PFaithBench.
Comments22 pages
