Text2VLM: Adapting Text-Only Datasets to Evaluate Alignment Training in Visual Language Models

Gabriel Downer, Sean Craven, Damian Ruck +1

The increasing integration of Visual Language Models (VLMs) into AI systems necessitates robust model alignment, especially when handling multimodal content that combines text and…