arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.32665cs.AIcs.LG
收藏

语言模型是否信任自身的解释?一项关于功能一致性的研究

Do Language Models Trust Their Own Justifications? A Study on Functional Consistency

Alisson Rosa Pereira, Bruno Iochins Grisci

AI总结:

本研究提出基于充分性、全面性、反事实性的实验框架,评估闭源及开源LLM自解释的功能一致性,发现不同LLM的解释可靠性存在差异。

摘要:

大语言模型(LLM)已被广泛应用于文本分类任务中,它们不仅输出类别预测,还会生成解释,突出显示被认为对得出预测标签最相关的词元。然而,这些被突出的元素在针对性干预下是否与模型自身的预测在行为上一致,这一点仍不明确,我们将此称为功能自一致性。尽管大量文献评估了此类解释的文本合理性,但很少有研究评估它们与模型实际行为的功能一致性。在本研究中,我们基于自一致性原则提出了一个实验框架:如果模型将某些词元识别为决定性因素,那么对其进行隔离、移除或语义反转,应会在模型的预测中产生系统性且可解释的变化。我们通过充分性、 comprehensiveness(全面性)和反事实性指标来实施此评估,并在IMDB和Steam评论上对闭源LLM(GPT-4o-mini)以及开源LLM(Gemma3、Granite8B、DeepSeek)进行了实验。结果显示,GPT-4o-mini在所有指标上都遵循预期的变化过程;Gemma3和Granite8B在充分性下保持一致性,但在要求更高的干预下失去一致性;而DeepSeek变体则表现出结构偏差,要么无法保持充分性,要么在全面性和反事实性下反应过度。这些发现表明,解释的可靠性因LLM家族和规模而异,较小的模型表现出矛盾,而较大的模型则表现出过度敏感性。通过结合充分性、全面性和反事实性,我们的方法为评估LLM自解释的功能一致性提供了一种系统的方法。

英文摘要:

Large Language Models (LLMs) have been widely adopted in text classification tasks, where they not only output class predictions but also generate explanations that highlight the tokens deemed most relevant for reaching the predicted label. Yet it remains unclear whether these highlighted elements are behaviorally consistent with the model's own predictions under targeted interventions, which we refer to as functional auto-consistency. While much of the literature evaluates the textual plausibility of such explanations, few studies assess their functional consistency with the model's actual behavior. In this work, we propose an experimental framework based on the principle of auto-consistency: if a model identifies certain tokens as decisive, then isolating, removing, or semantically inverting them should produce systematic and interpretable changes in its predictions. We operationalize this evaluation through sufficiency, comprehensiveness, and counterfactuality metrics, and conduct experiments on IMDB and Steam reviews across both closed-source (GPT-4o-mini) and open-source LLMs (Gemma3, Granite8B, DeepSeek). Results show that GPT-4o-mini follows the expected progression across all metrics, Gemma3 and Granite8B maintain coherence under sufficiency but lose consistency under more demanding interventions, while DeepSeek variants display structural deviations, either failing to preserve sufficiency or overreacting under comprehensiveness and counterfactuality. These findings show that explanation reliability varies across LLM families and scales, with smaller models displaying contradictions and larger ones exhibiting over-sensitivity. By combining sufficiency, comprehensiveness, and counterfactuality, our approach provides a systematic methodology for assessing the functional consistency of LLM self-explanations.

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。

↑