通过跨模态信息流解释和增强大型视觉语言模型中的情感回路

Computer Science > Computer Vision and Pattern Recognition arXiv:2605.21980 (cs) [Submitted on 21 May 2026 (v1), last revised 24 Sep 2026 (this version, v2)] Title:Interpreting and Enhancing Emotional Circuits in Large Vision-Language Models via Cross-Modal Information Flow Authors:Chengsheng Zhang, Chenghao Sun, Zhining Xie, Xinmei Tian View a PDF of the paper titled Interpreting and Enhancing Emotional Circuits in Large Vision-Language Models via Cross-Modal Information Flow, by Chengsheng Zhang and Chenghao Sun and Zhining Xie and Xinmei Tian View PDF HTML (experimental) Abstract:Large Vision-Language Models (LVLMs) represent a significant leap towards empathetic agents, demonstrating remarkable capabilities in emotion understanding. However, the internal mechanisms governing how LVLMs translate abstract visual stimuli into coherent emotional narratives remain largely unexplored, primarily due to the scarcity of visual counterfactuals and the diffuse nature of emotional expression. In this paper, we bridge this gap by introducing a steering-vector-based causal attribution framework tailored for descriptive emotional reasoning. To this end, we construct a specialized dataset to demystify the emotional circuits underlying the three-stage ``Adapt-Aggregate-Execute'' mechanism. Crucially, we discover a functional decoupling: visual emotional cues are aggregated in middle layers via sentiment-specific attention heads, but are subsequently translated into narrative generation in deep layers through emotion-general pathways. Guided by these insights, we regulate the emotional information routing to strengthen attention flow and amplify the semantic activation to consolidate expression. Extensive experiments on the comprehensive MER-UniBench demonstrate that our methods significantly improve performance via inference-time intervention, effectively mitigating emotional hallucinations and corroborating the causal fidelity of the discovered circuits. Comments: Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2605.21980 [cs.CV] (or arXiv:2605.21980v2 [cs.CV] for this version) https://doi.org/10.48550/arXiv.2605.21980 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Chengsheng Zhang [view email] [v1] Thu, 21 May 2026 04:20:12 UTC (6,739 KB) [v2] Thu, 24 Sep 2026 09:16:34 UTC (5,789 KB) Full-text links: Access Paper: View a PDF of the paper titled Interpreting and Enhancing Emotional Circuits in Large Vision-Language Models via Cross-Modal Information Flow, by Chengsheng Zhang and Chenghao Sun and Zhining Xie and Xinmei Tian View PDF HTML (experimental) TeX Source view license Current browse context: cs.CV new | recent | 2026-05 Change to browse by: cs cs.AI References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)