Search papers, labs, and topics across Lattice.
This paper introduces MMOOC, a comprehensive benchmark designed to evaluate the performance of Multimodal Large Language Models (MLLMs) in handling out-of-context (OOC) and shifted in-context (Shifted IC) questions. By incorporating over 41,000 image-question pairs that include both answerable Shifted IC cases and unanswerable OOC cases, the benchmark addresses significant gaps in existing evaluations that overlook nuanced context shifts. Experimental results reveal that current MLLMs struggle to effectively balance answering and refusal rates in the presence of context shifts, highlighting critical areas for improvement in model robustness.
Current MLLMs falter under context shifts, with a notable inability to balance answering and refusal rates, as revealed by the new MMOOC benchmark.
Multimodal Large Language Models (MLLMs) have achieved strong performance on a wide range of vision-language tasks, but often fail under imperfect or shifted contexts. A reliable MLLM should refuse truly out-of-context (OOC) questions with subject-level context shifts while still answering shifted in-context (Shifted IC) questions with non-subject context shifts. Existing benchmarks mainly target OOC or visually unanswerable questions, but overlook answerable Shifted IC cases and cover limited OOC shifts. To fill this gap, we present MMOOC, a large-scale benchmark for evaluating refusal and robust answering abilities of MLLMs. MMOOC contains over 41K image-question pairs, including answerable Shifted IC cases and unanswerable OOC cases, spanning three question formats, eight shift types and six visual scenarios, with data quality ensured through MLLM-based filtering and human verification. We evaluate model responses using Accuracy and Refusal Rate, and further introduce an LLM-as-a-Judge metric to assess the correctness of model reasoning. Experiments on diverse MLLMs show that current models still struggle to balance answer-ability and refusal under shifted contexts. We further analyze key failure patterns and show that post-training can improve robustness. MMOOC will be made publicly available.
Stanford HAI