1 paper · 1 filter
Xixi Liu, Ailin Deng, Christopher Zach
Mitigating object hallucination in large vision-language models (LVLMs) is critical to their safe deployment. Existing methods either are restricted to specific decoding methods, o…