DeP mitigates MLLM hallucinations by dynamically perturbing text prompts to identify and reinforce stable visual evidence regions while counteracting language prior biases using attention variance and logit statistics.
Advances in Neural Information Processing Systems37, 122811–122832 (2024)
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
ADAPT reduces MLLM hallucinations 40-60% by aligning cross-attention dynamics via visual anchors, supervised inference, and preference tuning while preserving general capabilities.
citing papers explorer
-
Decoding by Perturbation: Mitigating MLLM Hallucinations via Dynamic Textual Perturbation
DeP mitigates MLLM hallucinations by dynamically perturbing text prompts to identify and reinforce stable visual evidence regions while counteracting language prior biases using attention variance and logit statistics.
-
ADAPT: Attention Dynamics Alignment with Preference Tuning for Faithful MLLMs
ADAPT reduces MLLM hallucinations 40-60% by aligning cross-attention dynamics via visual anchors, supervised inference, and preference tuning while preserving general capabilities.