This paper explores the application of instruction tuning to enhance the capabilities of large language models (LLMs) in information retrieval (IR) tasks. The authors introduce INTERS, a novel instruction tuning dataset that covers 20 tasks across three fundamental IR categories: query understanding, document understanding, and query-document relationship understanding. The dataset is derived from 43 distinct datasets with manually crafted templates. Empirical results show that INTERS significantly improves the performance of various LLMs, such as LLaMA, Mistral, and Falcon, in IR tasks. The study also conducts extensive experiments to analyze the effects of instruction design, template diversity, few-shot demonstrations, and the volume of instructions on performance. The authors conclude that INTERS effectively enhances LLMs' proficiency in IR tasks and encourages further research in this area.This paper explores the application of instruction tuning to enhance the capabilities of large language models (LLMs) in information retrieval (IR) tasks. The authors introduce INTERS, a novel instruction tuning dataset that covers 20 tasks across three fundamental IR categories: query understanding, document understanding, and query-document relationship understanding. The dataset is derived from 43 distinct datasets with manually crafted templates. Empirical results show that INTERS significantly improves the performance of various LLMs, such as LLaMA, Mistral, and Falcon, in IR tasks. The study also conducts extensive experiments to analyze the effects of instruction design, template diversity, few-shot demonstrations, and the volume of instructions on performance. The authors conclude that INTERS effectively enhances LLMs' proficiency in IR tasks and encourages further research in this area.