Recent research has focused on advancements in out-of-distribution (OOD) detection, particularly in the context of computer vision and language models. A paper titled 'A Noisy Elephant in the Room' explores the effectiveness of computer vision systems in identifying unfamiliar images when training labels are unreliable. Another study, 'HGOE: Hybrid External and Internal Graph Outlier Exposure for Graph Out-of-Distribution Detection,' discusses methods for enhancing OOD detection in graph-based systems. A survey titled 'Generalized Out-of-Distribution Detection and Beyond in Vision Language Model Era' reviews the evolution and future challenges of OOD detection in vision language models. Additionally, Tsinghua University has made strides in improving text embeddings for smaller language models, such as MiniCPM, Phi-2, and Gemma, through contrastive fine-tuning, achieving an average performance gain of 56.33% across various benchmarks. The research highlights the growing importance of OOD detection techniques in both computer vision and natural language processing.