VIF is a new inference-time module that maintains visual grounding in MLLMs by directly bridging pure visual representations to the output space throughout generation.
Where is the woman’s blue bag located in the image?
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Vision Inference Former: Sustaining Visual Consistency in Multimodal Large Language Models
VIF is a new inference-time module that maintains visual grounding in MLLMs by directly bridging pure visual representations to the output space throughout generation.