Disrupted Representations

该数据集深入研究了在潜在扩散模型的微调过程中，操纵潜在分布所产生的影响。此外，该数据集重点关注评估不同防御目标对潜在分布的影响。该任务旨在针对潜在扩散模型进行数据保护。

This dataset conducts an in-depth investigation into the impacts of manipulating the latent distribution during the fine-tuning process of Latent Diffusion Models (LDMs). Additionally, it focuses on evaluating the impacts of different defensive objectives on the latent distribution. The core objective of this task is to enable data protection for latent diffusion models.