arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.04550cs.AI
收藏

大型语言模型中的推理鲁棒性基准测试

Benchmarking Reasoning Robustness in Large Language Models

Wuhan University(武汉大学);Nanyang Technological University(南洋理工大学)

Tong Yu, Yongcheng Jing, Xikun Zhang, Wentao Jiang, Wenjie Wu, Yingjie Wang, Wenbin Hu, Bo Du, Dacheng Tao

AI总结:

本文首次揭示了LLMs在推理鲁棒性上的四大局限,并提出Math-RoB基准,利用缺失信息触发的幻觉生成多样化数据集,以全面评估并推动更鲁棒推理框架的发展。

摘要:

尽管大型语言模型(LLMs)如DeepSeek在推理方面近期取得了成功,我们首次识别出推理鲁棒性和泛化能力中的一个关键困境:在新型或不完整数据上性能显著下降,这表明模型依赖于记忆模式而非系统性推理。我们的深入检查揭示了该问题背后的四个关键独特局限:(1)位置偏差——模型在多查询输入中偏爱较早的查询,但在后者中回答错误(例如,GPT-4o的准确率从75.8%降至72.8%);(2)指令敏感性——在Qwen2.5系列中性能下降5.0至7.5%,在DeepSeek-V3中下降5.0%(带有辅助指导);(3)数值脆弱性——数值替换急剧降低准确率(例如,GPT-4o从97.5%降至82.5%,GPT-o1-mini从97.5%降至92.5%);以及(4)记忆依赖——模型在缺失关键数据时诉诸猜测。这些发现进一步凸显了对启发式回忆而非严格逻辑推理的依赖,展示了推理鲁棒性方面的挑战。为全面调查这些鲁棒性挑战,本文引入了一个新颖的基准,称为Math-RoB,利用缺失信息触发的幻觉来暴露推理差距。这通过基于指令的方法生成紧密类似训练分布的多样化数据集来实现,促进整体鲁棒性评估并推动更鲁棒推理框架的发展。

英文摘要:

Despite the recent success of large language models (LLMs) in reasoning such as DeepSeek, we for the first time identify a key dilemma in reasoning robustness and generalization: significant performance degradation on novel or incomplete data, suggesting a reliance on memorized patterns rather than systematic reasoning. Our closer examination reveals four key unique limitations underlying this issue:(1) Positional bias--models favor earlier queries in multi-query inputs but answering the wrong one in the latter (e.g., GPT-4o's accuracy drops from 75.8 percent to 72.8 percent); (2) Instruction sensitivity--performance declines by 5.0 to 7.5 percent in the Qwen2.5 Series and by 5.0 percent in DeepSeek-V3 with auxiliary guidance; (3) Numerical fragility--value substitution sharply reduces accuracy (e.g., GPT-4o drops from 97.5 percent to 82.5 percent, GPT-o1-mini drops from 97.5 percent to 92.5 percent); and (4) Memory dependence--models resort to guesswork when missing critical data. These findings further highlight the reliance on heuristic recall over rigorous logical inference, demonstrating challenges in reasoning robustness. To comprehensively investigate these robustness challenges, this paper introduces a novel benchmark, termed as Math-RoB, that exploits hallucinations triggered by missing information to expose reasoning gaps. This is achieved by an instruction-based approach to generate diverse datasets that closely resemble training distributions, facilitating a holistic robustness assessment and advancing the development of more robust reasoning frameworks. Bad character(s) in field Abstract.

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。

↑