When Model Priors Conflict with Visual Evidence: Mitigating Commonsense-Driven Hallucinations by Selective Prior Calibration 文章

ArXiv CS.CV2026-08-03PAPERen作者: Kesheng Chen, Yamin Hu, Wenjian Luo

详细信息

来源站点
ArXiv CS.CV
作者
Kesheng Chen, Yamin Hu, Wenjian Luo
文章类型
PAPER
语言
en
发布日期
2026-08-03

摘要

arXiv:2607.29240v1 Announce Type: new Abstract: In vision--language models, commonsense-driven hallucination (CDH) occurs when a model's commonsense prior overrides clear visual evidence of an atypical state. For example, a model may report that a visibly six-fingered hand has five fingers. We show that these errors are systematically directed: when a model answers a question about a counterfactual (CF) image incorrectly, its answer often coincides with the candidate it prefers without access to the image. Suppressing this prior indiscriminately can repair CF errors, but may also disrupt correct answers on matched commonsense (CS) images, where the same prior is helpful. We therefore propose Selective Prior Calibration (SPC), which subtracts candidate-level prior-preference estimates from image-conditioned scores with an instance-dependent strength and revises the original prediction only when the resulting score pattern strongly supports an alternative.

相关事件

暂无数据

相关公司查看全部 (4)

A
ANDINONPROFIT
A
AT TCOMPANY
A
ActuaNONPROFIT

相关人物

暂无数据