Abstract
We investigate the internal representations of vision-language models (VLMs) to address hallucinations, a persistent challenge despite advances in model size and training. We project VLMs' internal image representations to their language vocabulary and observe more confident output probabilities on real objects than hallucinated objects. We additionally use these output probabilities to spatially localize real objects. Building on this approach, we introduce a knowledge erasure algorithm that removes hallucinations by linearly orthogonalizing image features with respect to hallucinated object features. We show that targeted edits to a model's latent representations can reduce hallucinations by up to 25.7% on the COCO2014 dataset while preserving performance. Our findings demonstrate how a deeper understanding of VLMs' latent representations can enhance reliability and enable novel capabilities, such as zero-shot segmentation.
Abstract (translated)
我们研究了视觉语言模型(VLMs)的内部表示,以解决在模型大小和训练方面取得进展但仍然存在的一种普遍挑战:幻觉。我们将VLMs的内部图像表示投影到其语言词汇中,并观察到真实物体上的输出概率比幻觉物体上的输出概率更自信。此外,我们还使用这些输出概率将真实物体进行空间局部化。在此基础上,我们引入了一种知识消逝算法,通过将图像特征与幻觉物体特征之间进行线性正交操作来消除幻觉。我们在COCO2014数据集上展示了针对模型 latent 表示的定向修改可以减少幻觉,同时保持性能。我们的研究结果表明,对VLMs latent 表示的更深入了解可以提高可靠性并实现诸如零 shot分割等新颖功能。
URL
https://arxiv.org/abs/2410.02762