待翻译：Taming Outlier Tokens in Diffusion Transformers

AI 服务暂时不可用，以下为来源正文，待恢复后补全翻译。

content type paperpublished August 2026 Taming Outlier Tokens in Diffusion Transformers AuthorsXiaoyu Wu†*, Yifei Wang†*, Tsu-Jui Fu, Liang-Chieh Chen, Zhe Gan, Chen Wei† View publication We study outlier tokens in Diffusion Transformers (DiTs) for image generation. Prior work has shown that Vision Transformers (ViTs) can produce a small number of high-norm tokens that attract disproportionate attention while carrying limited local information, but their role in generative models remains underexplored. We show that this phenomenon appears in both the encoder and denoiser of modern Representation Autoencoder (RAE)-DiT pipelines: pretrained ViT encoders can produce outlier representations, and DiTs themselves can develop internal outlier tokens, especially in intermediate layers. Moreover, simply masking high-norm tokens does not improve performance, indicating that the problem is not only caused by a few extreme values, but is more closely related to corrupted local patch semantics. To address this issue, we introduce Dual-Stage Registers (DSR), a register-based intervention for both components: trained registers when available, recursive test-time registers otherwise, and diffusion registers for the denoiser. Across ImageNet and large-scale text-to-image generation, these interventions consistently reduce outlier artifacts and improve generation quality. Our results highlight outlier-token control as an important ingredient in building stronger DiTs. † Rice University * Equal contribution DiT-Air: Revisiting the Efficiency of Diffusion Model Architecture Design in Text to Image Generation December 11, 2025research area Computer Vision In this work, we empirically study Diffusion Transformers (DiTs) for text-to-image generation, focusing on architectural choices, text-conditioning strategies, and training protocols. We evaluate a range of DiT-based architectures—including PixArt-style and MMDiT variants—and compare them with a standard DiT variant which directly processes concatenated text and noise inputs. Surprisingly, our findings reveal that the performance of standard… Read more On Inductive Biases That Enable Generalization of Diffusion Transformers September 22, 2025research area Computer Visionconference NeurIPS Recent work studying the generalization of diffusion models with UNet-based denoisers reveals inductive biases that can be expressed via geometry-adaptive harmonic bases. However, in practice, more recent denoising networks are often based on transformers, e.g., the diffusion transformer (DiT). This raises the question: do transformer-based denoising networks exhibit inductive biases that can also be expressed via geometry-adaptive harmonic… Read more