Please cite this work with the following BibTeX: @inproceedings{cocchi2024augmenting, title={{Augmenting Multimodal LLMs with Self-Reflective Tokens for Knowledge-based Visual Question Answering}}, ...