COPSD-AfriMGSM: Multilingual Math Reasoning Training Data via Crosslingual Distillation

A training dataset associated with the paper 'Crosslingual On-Policy Self-Distillation for Multilingual Reasoning'. It is designed to improve mathematical reasoning in low-resource languages using the COPSD method. The dataset was created by yihongLiu and was last updated on Hugging Face in May 2026.

Use Cases

• Training multilingual language models for mathematical reasoning based on the described crosslingual distillation method.
• Benchmarking model performance on low-resource language tasks based on the dataset's stated purpose.
• Studying knowledge transfer from high-resource to low-resource languages based on the COPSD framework mentioned in the description.

Strengths

• Dataset is explicitly linked to a published research paper on crosslingual reasoning.
• Methodology is clearly stated as Crosslingual On-Policy Self-Distillation (COPSD).
• Last update timestamp is precisely recorded as 2026 05 13 21:08:13.

Limitations

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

Provenance

Source

Hugging Face dataset repository by author yihongLiu.

Collection Method

Created for the paper 'Crosslingual On-Policy Self-Distillation for Multilingual Reasoning'; method involves transferring reasoning behavior from high-resource to low-resource languages.

Time Range

null

Freshness

Last updated 2026-05-13 21:08:13; freshness should be verified.

Geography

null