Work

Spotlight and Shadow: Attention-Guided Dual-Anchor Introspective Decoding for MLLM Hallucination Mitigation

Year: 2026

Type: conference-paper

Abstract: Multimodal Large Language Models (MLLMs) have demonstrated remarkable reasoning capabilities yet continue to suffer from hallucination, where generated text contradicts visual content.In this paper, we introduce Dual-Anchor Introspective Decoding (DAID), a novel contrastive decoding framework that dynamically calibrates each token generation by mining the model's internal perceptual discrepancies.Specifically, DAID identifies a Spotlight layer to amplify visual factual signals and a Shadow layer to suppress textual inertia.By leveraging visual attention distributions to guide this dual-anchor selection process, our method ensures precise, tokenspecific adaptation.Experimental results across multiple benchmarks and MLLMs demonstrate that DAID significantly mitigates hallucination while enhancing general reasoning capabilities.

Source: Findings of the Association for Computational Linguistics: ACL (Findings of ACL)

DOI: 10.18653/v1/2026.findings-acl.646

Authors: Yebo Wu, Han Jin, Zhijiang Guo, Li Li

Institutions: University of Macau, Hong Kong University of Science and Technology, The Hong Kong University of Science and Technology (Guangzhou)

Language: English

Topic: Multimodal Machine Learning Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Funders: United Mitochondrial Disease Foundation, Universidade de Macau

Award: G180185047