Minyeong Kim

2

papers

13

total citations

papers (2)

Exploiting Semantic Reconstruction to Mitigate Hallucinations in Vision-Language Models

HalLoc: Token-level Localization of Hallucinations for Vision Language Models