Diverse Demonstrations Improve In-context Compositional Generalization We empirically show that combining diverse demonstrations with in-context learning substantially improves performance across three compositional generalization semantic parsing datasets in the pure in-context learning setup and when combined with finetuning
Diverse Demonstrations Improve In-context Compositional Generalization . . . We empirically show that combining diverse demonstrations with in-context learning substantially improves performance across three compositional generalization semantic parsing datasets in the pure in-context learning setup and when combined with finetuning
Ben Bogin - Google Scholar Diverse Demonstrations Improve In-context Compositional Generalization I Levy*, B Bogin*, J Berant Proceedings of the 61st Annual Meeting of the Association for Computational …
Diverse Demonstrations Improve In-context Compositional Generalization We empirically show that combining diverse demonstrations with in-context learning substantially improves performance across three compositional generalization semantic parsing datasets in the pure in-context learning setup and when combined with finetuning
In-Context Compositional Generalization for Large Vision-Language Models In this paper, we consider the asymmetry charac-teristics of multimodal data from both the content and structure perspectives, and propose a demon-stration selection method for LVLMs to achieve in-context compositional generalization
Diverse Demonstrations Improve In-context Compositional Generalization . . . We empirically show that combining diverse demonstrations with in-context learning substantially improves performance across three compositional generalization semantic parsing datasets in the pure in-context learning setup and when combined with finetuning