HSSBench: Benchmarking Humanities and Social Sciences Ability for Multimodal Large Language Models
This addresses a gap in benchmarking for HSS tasks, which require interdisciplinary thinking, but it is incremental as it focuses on creating a new benchmark rather than advancing model capabilities directly.
The authors tackled the lack of benchmarks for evaluating multimodal large language models (MLLMs) in humanities and social sciences (HSS) by introducing HSSBench, a dataset with over 13,000 samples in multiple languages, and found that it poses significant challenges for over 20 state-of-the-art models.
Multimodal Large Language Models (MLLMs) have demonstrated significant potential to advance a broad range of domains. However, current benchmarks for evaluating MLLMs primarily emphasize general knowledge and vertical step-by-step reasoning typical of STEM disciplines, while overlooking the distinct needs and potential of the Humanities and Social Sciences (HSS). Tasks in the HSS domain require more horizontal, interdisciplinary thinking and a deep integration of knowledge across related fields, which presents unique challenges for MLLMs, particularly in linking abstract concepts with corresponding visual representations. Addressing this gap, we present HSSBench, a dedicated benchmark designed to assess the capabilities of MLLMs on HSS tasks in multiple languages, including the six official languages of the United Nations. We also introduce a novel data generation pipeline tailored for HSS scenarios, in which multiple domain experts and automated agents collaborate to generate and iteratively refine each sample. HSSBench contains over 13,000 meticulously designed samples, covering six key categories. We benchmark more than 20 mainstream MLLMs on HSSBench and demonstrate that it poses significant challenges even for state-of-the-art models. We hope that this benchmark will inspire further research into enhancing the cross-disciplinary reasoning abilities of MLLMs, especially their capacity to internalize and connect knowledge across fields.