K-LLaVA-W: Korean Vision-Language Model Evaluation Benchmark

K-LLaVA-W is a Korean adaptation of the LLaVA-Bench-in-the-wild, designed for evaluating vision-language models. The benchmark was created by translating the original English text into Korean and reviewing its naturalness through human inspection. It was published by NCSOFT and last updated on July 25, 2025.

Use Cases

• Benchmarking Korean vision-language model performance based on translated and localized image-text pairs.
• Evaluating model understanding of Korean text within visual contexts.
• Assessing the naturalness and cultural relevance of multimodal outputs for Korean language applications.

Strengths

• Human-reviewed translation for naturalness in Korean.
• Specifically localized by changing English text in images to Korean.
• Published by a known organization, NCSOFT.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

NCSOFT via Hugging Face.

Collection Method

Translated and localized from the LLaVA-Bench-in-the-wild.

Time Range

null

Freshness

Last updated 2025-07-25 06:37:21; freshness should be verified.

Geography

null