📄 New paper MemLens: Benchmarking Multimodal Long-Term Memory in Large Vision-Language Models released on arXiv.