concept-binding

This paper introduces M2BIND, a benchmark to evaluate whether vision-language models maintain stable visual-linguistic associations across languages. It finds that binding is not language-invariant, with cross-family and cross-script settings causing significant performance collapse and weaker internal causal binding.

本文探讨了CLIP为何在概念绑定上表现不佳，表明虽然CLIP的绑定函数复杂度高，但受控的Transformer模型可以通过乘法交互学习复杂度较低的绑定函数，从而更好地泛化。