Culture Mt

1,002 user-generated notes (UGC) spanning 14 content domains form the CULTURE-MT benchmark for evaluating cultural transmission and emotion resonance in translation. It was created by Wulinjuan and presented at ICML 2026. The dataset focuses on Chinese-to-English social media translation beyond literal meaning.

Use Cases

• Benchmarking machine translation models for cultural effectiveness based on the described cultural transmission metric.
• Evaluating emotion resonance in translated social media content based on the UGC-specific emotion resonance metric.
• Training or fine-tuning translation models for social media text based on the 14 described content domains.
• Researching cultural adaptation strategies in cross-lingual natural language processing based on the described cultural effectiveness focus.

Strengths

• Contains 1,002 specific user-generated notes, providing a defined scale for evaluation.
• Spans 14 distinct content domains, suggesting thematic diversity.
• Specifically designed for evaluating cultural transmission and emotion resonance, providing a clear, specialized purpose.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but the size, file formats, and sample data are unknown, which may limit suitability assessment.
• Data may reflect geographic and platform bias inherent to its source of Chinese social media UGC.

Provenance

Source

Author Wulinjuan, hosted on Hugging Face.

Collection Method

Likely curated from Chinese social media user-generated content, though the exact gathering method is unspecified.

Freshness

Last updated 2026-05-28 00:40:43; freshness should be verified.

Geography

Primarily Chinese-language source content, with English translation targets.