On the Performance of Multimodal Language Models

Utsav Garg, Erhan Bas

Instruction-tuned large language models (LLMs) have demonstrated promising zero-shot generalization capabilities across various downstream tasks. Recent research has introduced mul…