Uncovering Grounding IDs: How External Cues Shape Multi-Modal Binding
This work addresses interpretability and robustness issues in multimodal AI for researchers and practitioners, though it is incremental as it builds on prior findings about visual structures.
The paper tackled the problem of limited structured reasoning and precise grounding in large vision-language models by investigating how external visual cues improve multimodal binding, finding that Grounding IDs emerge as latent identifiers that enhance cross-modal alignment and reduce hallucinations.
Large vision-language models (LVLMs) show strong performance across multimodal benchmarks but remain limited in structured reasoning and precise grounding. Recent work has demonstrated that adding simple visual structures, such as partitions and annotations, improves accuracy, yet the internal mechanisms underlying these gains remain unclear. We investigate this phenomenon and propose the concept of Grounding IDs, latent identifiers induced by external cues that bind objects to their designated partitions across modalities. Through representation analysis, we find that these identifiers emerge as robust within-partition alignment in embedding space and reduce the modality gap between image and text. Causal interventions further confirm that these identifiers mediate binding between objects and symbolic cues. We show that Grounding IDs strengthen attention between related components, which in turn improves cross-modal grounding and reduces hallucinations. Taken together, our results identify Grounding IDs as a key symbolic mechanism explaining how external cues enhance multimodal binding, offering both interpretability and practical improvements in robustness.