JAMMEval: A Refined Collection of Japanese Benchmarks for Reliable VLM Evaluation
This work addresses the need for more reliable benchmarks in Japanese VQA to support accurate model comparisons, though it is incremental as it refines existing datasets rather than introducing a new paradigm.
The paper tackled the problem of unreliable evaluation for Japanese vision-language models (VLMs) due to issues like ambiguous questions and incorrect answers in existing benchmarks, resulting in JAMMEval, a refined collection that improves data quality and yields evaluation scores with lower variance and better model distinction.
Reliable evaluation is essential for the development of vision-language models (VLMs). However, Japanese VQA benchmarks have undergone far less iterative refinement than their English counterparts. As a result, many existing benchmarks contain issues such as ambiguous questions, incorrect answers, and instances that can be solved without visual grounding, undermining evaluation reliability and leading to misleading conclusions in model comparisons. To address these limitations, we introduce JAMMEval, a refined collection of Japanese benchmarks for reliable VLM evaluation. It is constructed by systematically refining seven existing Japanese benchmark datasets through two rounds of human annotation, improving both data quality and evaluation reliability. In our experiments, we evaluate open-weight and proprietary VLMs on JAMMEval and analyze the capabilities of recent models on Japanese VQA. We further demonstrate the effectiveness of our refinement by showing that the resulting benchmarks yield evaluation scores that better reflect model capability, exhibit lower run-to-run variance, and improve the ability to distinguish between models of different capability levels. We release our dataset and code to advance reliable evaluation of VLMs.