测试大模型对遮蔽文本的推理能力，发现其表现严重依赖语义线索。
Exploring the Limits of Large Language Models: A Systematic Evaluation of Masked Text Processing Ability through MskQA and MskCal
- 设计新任务MskQA和MskCal，评估模型在遮蔽文本下的问答与计算能力
- 全遮蔽时性能大幅下降，说明模型依赖表面模式而非深层理解
- GPT-4o比4o-mini更强，尤其在数值推理任务中表现更优
本文通过系统评估大语言模型（LLMs）处理遮蔽文本的能力，揭示其局限性。提出两个新任务：MskQA，用于衡量在如RealtimeQA等数据集上的遮蔽问答推理能力；MskCal，评估遮蔽算术问题中的数值推理能力。测试GPT-4o与4o-mini发现，尽管模型对遮蔽具有一定韧性，但性能高度依赖遮蔽率与语义线索。当语义线索完全缺失（‘solid masking’）时，性能显著低于部分线索保留（‘partial lifting’），表明模型依赖表层模式。有趣的是，GPT-4o始终优于4o-mini，尤其在MskCal任务中表现出更强的数值推理能力。研究揭示了背景知识与推理能力在遮蔽文本处理中的交互关系，推动对大模型真实理解能力的深入认知，并强调需发展更鲁棒的评估方法。
原文摘要 · Abstract (English)
This paper sheds light on the limitations of Large Language Models (LLMs) by rigorously evaluating their ability to process masked text. We introduce two novel tasks: MskQA, measuring reasoning on masked question-answering datasets like RealtimeQA, and MskCal, assessing numerical reasoning on masked arithmetic problems.Testing GPT-4o and 4o-mini reveals that while LLMs exhibit some resilience to masked text, their performance is highly contingent on masking rates and semantic cues. Specifically, "solid masking," where semantic clues are entirely absent, leads to a significant performance drop compared to "partial lifting," where some semantic information is retained, indicating LLMs' reliance on surface-level patterns. Interestingly, GPT-4o consistently outperforms 4o-mini, particularly in MskCal, demonstrating a greater ability to handle numerical reasoning with masked text. This underscores the crucial role of semantic cues in the reasoning process of LLMs. Our study illuminates the interplay between background knowledge and reasoning ability in masked text processing, paving the way for a deeper understanding of LLM capabilities and limitations, and highlighting the need for more robust evaluation methods to accurately assess their true comprehension abilities.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。