Fixing Semantic Blind Spots in Anchor Tokens of dMLLMs

Recent advances in diffusion-based Multimodal Large Language Models (dMLLMs) offer a compelling alternative to autoregressive counterparts; however, they remain prone to hallucinations. Through information flow analysis on representative dMLLMs (e.g., LLaDA-V and LaViDa), we identify two intertwined factors contributing to this issue. First, although the special image_newline tokens serve as semantic anchors for aggregating visual information, they simultaneously induce severe attention sinks, excessively consuming the model's attention budget. Second, the longrange decay inherent in Rotary Position Embedding (RoPE) leads to semantic blind spots, preventing these anchors from uniformly perceiving the entire visual input. Accordingly, our objective is to moderately alleviate the attention sink effect on semantic anchors while enhancing their ability to aggregate global visual information, thereby eliminating semantic blind spots. To this end, we propose Extrinsic Distance-Aware Regularization (EDAR), a training-free decoding strategy that augments the attention key space with a static, distanceaware matrix. This matrix jointly redistributes excessive attention away from anchors and injects absolute positional bias to ensure uniform visual coverage. Experiments on LLaDA-V and LaViDa demonstrate that EDAR effectively eliminates semantic blind spots and achieves state-of-the-art performance on both hallucination-specific and general multimodal benchmarks.