Thank you for open-sourcing your codebase for this awesome work. Regarding the Quiz evaluation, is it possible to share the prompt used to evaluate the model? Are the whole slide deck images given as input, or the Description of the slide are given as input, similar to PPTEval?