English(EN) Did I See It or Think It? Probing Reality Monitoring in Vision-Language Models with Concept Injection

Gemma-3-27B 在视觉语言模型中表现出脆弱的现实监控能力

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-09 05:00

研究人员开发了一种新方法来测试视觉语言模型（VLMs）的现实监控能力，特别是它们区分内部生成概念和实际图像内容的能力。研究发现，在测试的六个 VLM 中，只有 Gemma-3-27B 能够可靠地识别注入的概念，尽管其来源归因能力很脆弱。这种失败模式在普通字幕生成中也观察到，注入的对象出现在了相当比例的真实照片字幕中，这凸显了在评估多模态自我报告时，同时评估内容和报告来源的必要性。 AI

影响突显了当前视觉语言模型在区分内部想法与外部刺激方面的关键局限性，影响了它们在现实世界应用中的可靠性。

排序理由详细介绍视觉语言模型新评估方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CV 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 15 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍视觉语言模型新评估方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CV TIER_1 English(EN) · Tina Khezresmaeilzadeh, Seyedarmin Azizi, Mehdi Kamal, Massoud Pedram · 2026-11-09 05:00

是我看到了还是我想了？通过概念注入探究视觉语言模型的现实监控能力

arXiv:2611.09014v1 Announce Type: new Abstract: Recent work tests whether language models can introspect by injecting a concept direction into their activations and asking whether they notice it and can name it. These studies are text-only. Vision-language models (VLMs) face an a…

本文如何被排名

报道来源 [1]

是我看到了还是我想了？通过概念注入探究视觉语言模型的现实监控能力

相关实体

相关话题