大语言模型的显著性偏差：常识推理中的知识压制而非缺失

View PDF HTML (experimental)

Abstract:Despite advances in complex reasoning, large language models (LLMs) can prioritize explicit input conditions over implicit task prerequisites. In everyday commonsense reasoning, this can lead to a failure we term Salience Bias: salient but task-irrelevant details (e.g., numerical values) draw models into computation while they overlook the physical or commonsense prerequisites of the task. A critical open question is whether this failure reflects a genuine gap in commonsense knowledge or merely its suppression under misleading task framing. To investigate this, we construct the SaliTrap Benchmark, comprising 1,145 items in four trap dimensions. Evaluating 12 LLMs, we find substantial vulnerability across the tested models, with higher numerical distractor counts associated with lower trap-avoidance rates and trap recognition not always leading to avoidance. Further probing of sycophantic-compliance cases shows that the relevant commonsense can often be elicited when the original task framing is removed, suggesting a gap between recognizing a constraint and applying it during task execution. Building on this diagnosis, we further show that lightweight, inference-time prompting alone substantially closes the gap without any retraining. Our findings highlight the importance of applying commonsense constraints during task execution, and we release SaliTrap as a testbed for studying this gap. The codes are available at this https URL

Subjects:	Computation and Language (cs.CL)
Cite as:	arXiv:2607.28478 [cs.CL]
	(or arXiv:2607.28478v2 [cs.CL] for this version)
	https://doi.org/10.48550/arXiv.2607.28478 arXiv-issued DOI via DataCite

Submission history

From: Zheng Wu [view email]
[v1] Thu, 30 Jul 2026 16:30:08 UTC (2,441 KB)
[v2] Sun, 27 Sep 2026 02:32:33 UTC (2,997 KB)