Beyond Prompts: Unconditional 3D Inversion for Out-of-Distribution Shapes

ArXi:2604.14914v1 Announce Type: new Text-driven inversion of generative models is a core paradigm for manipulating 2D or 3D content, unlocking numerous applications such as text-based editing, style transfer, or inverse problems. However, it relies on the assumption that generative models remain sensitive to natural language prompts. We nstrate that for state-of-the-art native text-to-3D generative models, this assumption often collapses.