Two papers have been accepted accepted by ICLR 2025. One focuses on the compression of large language models (LLMs), and the other focuses on prompt-based expert routing for text-to-image diffusion models. One paper on depth pruning for LLMs has been accepted by NAACL 2025.