Aadarsh Anantha Ramakrishnan

CORDIAL: Can Multimodal Large Language Models Effectively Understand Coherence Relationships?

Multimodal Large Language Models (MLLMs) are renowned for their superior instruction-following and reasoning capabilities across diverse problem domains. However, existing benchmarks primarily focus on assessing factual and logical correctness in downstream tasks, with limited emphasis on evaluating…