On the Additive Compositionality of Task Vectors in Vision-Language Models.

Yuting Shi, Houjing Wei, Naoya Inoue

Browse the full EACL paper archive.