推理谜题：可解释性如何揭示视觉语言模型的认知局限
Reasoning Riddles: How Explainability Reveals Cognitive Limits in Vision-Language Models
AI总结：
该研究针对视觉语言模型在画谜这类横向思维任务中推理过程不透明的问题，构建了含221个画谜的标注数据集与分离推理质量和答案正确性的评估框架，通过三种提示策略的可解释性分析，揭示了模型的认知优势与局限，证明可解释性是模型性能的固有组成部分。
摘要：
视觉语言模型（Vision-Language Models, VLMs）在许多多模态任务中表现出色，但在画谜（rebus puzzles）这类复杂的横向思维挑战中，其认知过程仍不透明。尽管近期研究表明这些模型在画谜求解上存在明显困难，但其底层推理过程和失败模式在很大程度上仍未被探究。我们通过全面的可解释性分析填补这一空白，该分析超越性能指标，旨在理解VLM如何应对这些复杂的横向思维挑战。本研究贡献了一个包含221个画谜的系统标注数据集，覆盖六个认知类别，并配套了一个将推理质量与答案正确性分离的评估框架。我们研究了三种旨在激发不同类型解释过程的提示策略，揭示了关于VLM认知过程的关键洞见。研究结果表明，不同谜题类别的推理质量差异显著，模型在视觉组合方面表现出系统性优势，但在缺失解读和文化象征意义方面存在根本性局限。我们还发现，提示策略对认知方式和问题解决效果都有实质性影响，这确立了可解释性是模型性能的固有组成部分，而非事后考量。
英文摘要：
Vision-Language Models (VLMs) excel at many multimodal tasks, yet their cognitive processes remain opaque on complex lateral thinking challenges like rebus puzzles. While recent work has demonstrated these models struggle significantly with rebus puzzle solving, the underlying reasoning processes and failure patterns remain largely unexplored. We address this gap through a comprehensive explainability analysis that moves beyond performance metrics to understand how VLMs approach these complex lateral thinking challenges. Our study contributes a systematically annotated dataset of 221 rebus puzzles across six cognitive categories, paired with an evaluation framework that separates reasoning quality from answer correctness. We investigate three prompting strategies designed to elicit different types of explanatory processes and reveal critical insights into VLM cognitive processes. Our findings demonstrate that reasoning quality varies dramatically across puzzle categories, with models showing systematic strengths in visual composition while exhibiting fundamental limitations in absence interpretation and cultural symbolism. We also discover that prompting strategy substantially influences both cognitive approach and problem-solving effectiveness, establishing explainability as an integral component of model performance rather than a post-hoc consideration.
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。
