MixCuBe (Mixed Culture Benchmark) accompanies a 2025 paper evaluating cultural awareness in multimodal large language models. The dataset likely contains mixed-cultural inputs combining elements from multiple cultures, such as images and text. It was created by kyawyethu and published on Hugging Face.
Use Cases
- Benchmarking model cultural awareness based on mixed-cultural inputs
- Evaluating multimodal question-answering performance in cross-cultural contexts
- Testing image classification accuracy for culturally diverse visual elements
Strengths
- Dataset is directly linked to a specific research paper (WHEN TOM EATS KIMCHI)
- Platform tags indicate specific modalities (image, text) and tasks (question answering, image classification)
- Last updated on 2025-03-24, indicating recent maintenance
Limitations
- Column-level documentation is absent; field semantics must be inferred after download
- Row count and dataset size are unknown, which may limit suitability assessment
Provenance
- Source
- kyawyethu on Hugging Face
- Freshness
- Last updated 2025-03-24 12:54:34