The field of out-of-distribution (OOD) detection is moving towards more effective and efficient methods for distinguishing between in-distribution and out-of-distribution data. Recent research has focused on leveraging large language models, graph-based architectures, and self-supervised learning techniques to improve OOD detection performance. Notably, the use of positive and negative prompt supervision, test-time calibration, and inter-sample information has shown promising results. These innovative approaches have the potential to advance the field by addressing challenges such as semantically similar OOD samples, long-tailed distributions, and unreliable OOD detection. Noteworthy papers include: Out-of-Distribution Detection with Positive and Negative Prompt Supervision Using Large Language Models, which proposes a method that encourages negative prompts to capture inter-class features. Graph Out-of-Distribution Detection via Test-Time Calibration with Dual Dynamic Dictionaries, which introduces a novel test-time graph OOD detection method that calibrates OOD scores using dual dynamically updated dictionaries. BootOOD: Self-Supervised Out-of-Distribution Detection via Synthetic Sample Exposure under Neural Collapse, which presents a fully self-supervised OOD detection framework that bootstraps exclusively from ID data. Exploiting Inter-Sample Information for Long-tailed Out-of-Distribution Detection, which demonstrates that exploiting inter-sample relationships using a graph-based representation can significantly improve OOD detection in long-tailed recognition of vision datasets.