Autoregressive language models, when combined with In-Context Learning, have demonstrated strong performance across various tasks using limited labeled data. However, the application of this approach to Named Entity Recognition remains underexplored, partly due to the unique structural challenges inherent to the task. We propose a new technique for named entity recognition using autoregressive language models based on In-Context Learning and Information Retrieval techniques. Given an input text, models using this technique can fetch similar examples from a training set. These examples are then used to improve the language model’s ability to recognize named entities within the provided text. This approach is modular and independent of the underlying language model and the Information Retrieval technique. Experimental results indicate that this technique can be effective for low-resource applications, as in the CrossNER collection we achieve state-of-the-art performance with our proposed technique, showing that Information Retrieval can increase the F-score by as much as 11% points.