# RL Post-Training in Diffusion Transformers Is Activation Steering

> Source: <https://aiflash.com/news/155114/>
> Published: 2026-11-20 19:00:05+00:00

**Recommendation:** When RL post-training a diffusion transformer for text rendering or composition, treat the gains as activation steering and target prompt-dependent capabilities in your reward and data design.
More in the playbook "Fine-tuning and data": <https://wpnews.pro/playbooks/training-data>

Reinforcement learning (RL) post-training substantially improves the visual quality, text rendering, and compositional accuracy of diffusion transformers. Because capabilities such as rendering requested text or composing multiple objects are inherently prompt-dependent, a natural hypothesis is that
