hybrid-retrieval

What problem does it solve?

This Skill addresses the limitations of pure vector search in RAG systems by combining dense (semantic) and sparse (keyword) retrieval methods to improve the accuracy and relevance of search results.

Core Features & Use Cases

• Combines Dense and Sparse Retrieval: Integrates semantic understanding with keyword matching for more robust search.
• Handles Keyword-Dependent Queries: Effective when exact terms or specific keywords are crucial for accurate results.
• Improves RAG Performance: Enhances the quality of information retrieved for large language models.
• Use Case: When searching for technical documentation, a query like "troubleshoot error E-4521" might miss exact matches with pure vector search. Hybrid retrieval ensures that documents containing "E-4521" are prioritized alongside semantically related content.

Quick Start

Use the hybrid-retrieval skill to implement a search that combines dense and sparse retrieval methods for improved accuracy.