Culturay Kk V1: Kazakh-Language Texts from Internet Archive Corpus

Culturay Kk V1 is a Kazakh-language text dataset derived from the CulturaY multilingual corpus built from Internet Archive data. The subset underwent a multi-stage cleaning and quality-filtering pipeline. The dataset is prepared by author 'salyamq' and was last updated on June 13, 2026.

Use Cases

• Training Kazakh language models based on the described text corpus.
• Studying multilingual or code-switched text based on the mention of mixed Russian-Kazakh content.
• Developing text cleaning or filtering pipelines based on the described multi-stage processing.
• Analyzing web-derived text characteristics based on the source from Internet Archive data.

Strengths

• Derived from the CulturaY corpus, a large multilingual source built from Internet Archive data.
• Underwent a multi-stage cleaning and quality-filtering pipeline.
• The dataset is described as ready for SLM/LLM training.

Limitations

• The dataset is a first version and may contain mixed Russian-Kazakh text, minor spam, and adult content.
• 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

CulturaY multilingual corpus, built from Internet Archive data.

Collection Method

Derived and filtered from a larger corpus through a multi-stage cleaning pipeline.

Freshness

Last updated 2026-06-13 12:53:19; freshness should be verified.