GSM8K Cultural: Six Cultural Variants of a Math Word Problem Test Set

Six cultural variants of the GSM8K test set were created to investigate how cultural context influences large language model performance on mathematical problems. The dataset is part of a research project detailed in the paper 'Lost in Cultural Translation: Do LLMs Struggle with Math Across Cultural Contexts?' and was authored by abedk. It was last updated on Hugging Face in March 2025.

Use Cases

• Benchmarking LLM performance on math problems across different cultural contexts based on the six described variants.
• Analyzing cultural bias in LLM reasoning based on the dataset's methodology for creating cultural variants.
• Training or fine-tuning models for culturally-aware mathematical problem-solving based on the text-based problem sets.

Strengths

• Dataset is explicitly designed for a specific research purpose: investigating cultural context in LLM math performance.
• Methodology is documented in a referenced research paper and associated GitHub code.
• Last update was recorded on 2025-03-31, indicating recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset description is incomplete, requiring a visit to the external page for full details.

Provenance

Source

Hugging Face dataset authored by abedk, derived from the GSM8K test set.

Collection Method

Created as part of a research investigation; methodology detailed in the associated paper.

Time Range

null

Freshness

Last updated 2025-03-31 02:20:04

Geography

null