Decoding by Perturbation: Mitigating MLLM Hallucinations via Dynamic Textual Perturbation

arXiv:2604.12424v1 Announce Type: cross Abstract: Multimodal Large Language Models frequently suffer from inference hallucinations, partially stemming from language priors dominating visual evidence. Existing training-free mitigation methods either perturb the visual representation and deviate from the natural image distribution, or enforce intrusive manipulations that compromise the model's inherent generative fluency. We introduce a novel perspective that multimodal hallucination manifests as the hypersensitivity of visual grounding to textual phrasing during the decoding phase. Building on this insight, we propose Decoding by Perturbation (DeP), a training-free framework mitigating prior-induced hallucinations via controlled textual interventions. DeP employs a dynamic probe applying multi-level textual perturbations to elicit latent language priors. Leveraging attention variance, it enhances stable evidence regions while suppressing suspicious noise in the feature space. Furthermore, it constructs an interpretable prior drift direction using logits statistics to counteract probability biases from textual co-occurrences. Extensive experiments confirm DeP effectively reduces hallucinations and achieves superior performance across multiple benchmarks.

Related

• Faithful-First Reasoning, Planning, and Acting for Multimodal LLMs
• TARAC: Mitigating Hallucination in LVLMs via Temporal Attention Real-time Accumulative Connection
• 3D-VCD: Hallucination Mitigation in 3D-LLM Embodied Agents through Visual Contrastive Decoding
• Stabilizing Unsupervised Self-Evolution of MLLMs via Continuous Softened Retracing reSampling
• Variational Visual Question Answering for Uncertainty-Aware Selective Prediction

Source: arXiv cs.AI | 2026-04-15