Re-injecting shallow text features into deeper MMDiT blocks counteracts measured 'prompt forgetting' and improves instruction following in SD3, SD3.5, FLUX, and Qwen-Image without retraining.
arXiv preprint arXiv:2601.02211 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2representative citing papers
Early DC component convergence in text-to-image Transformer features causes output homogeneity; selective early attenuation via DAVE improves diversity without retraining or extra cost.
citing papers explorer
-
Prompt Reinjection: Alleviating Prompt Forgetting in Multimodal Diffusion Transformers for Text-to-Image Generation
Re-injecting shallow text features into deeper MMDiT blocks counteracts measured 'prompt forgetting' and improves instruction following in SD3, SD3.5, FLUX, and Qwen-Image without retraining.
-
Breaking the Lock-in: Diversifying Text-to-Image Generation via Representation Modulation
Early DC component convergence in text-to-image Transformer features causes output homogeneity; selective early attenuation via DAVE improves diversity without retraining or extra cost.