Turkish Law BGE-M3 Embeddings for Supreme Court Precedents and Legislation

Pre-computed embeddings for Turkish legal precedents from the Supreme Court and Council of State, as well as statutory legislation. The dataset is formatted in Parquet and optimized for Retrieval-Augmented Generation pipelines using a 'Multi-Index' architecture. It was created by muhamparlak and last updated on 2026-05-16.

Use Cases

• Semantic search for legal precedents based on the pre-computed embeddings.
• Enhancing Retrieval-Augmented Generation pipelines for Turkish legal Q&A systems.
• Analyzing similarity between statutory legislation and court decisions using the embedding vectors.

Strengths

• Data is specifically formatted and optimized for RAG pipelines.
• Covers two major sources of Turkish law: Supreme Court precedents and statutory legislation.
• Last updated on 2026-05-16, indicating potential recent maintenance.

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

Derived from Turkish legal precedents (Supreme Court/Yargıtay and Council of State/Danıştay) and statutory legislation.

Collection Method

Pre-computed embeddings, likely using the BGE-M3 model.

Freshness

Last updated 2026-05-16 19:51:12.

Geography

Turkey