Paper: RL post-training gains in diffusion transformers are mostly activation steering
Tuna Han Salih Meral · hf · 2026-11-21
A paper from Virginia Tech, fal, Georgia Tech and IBM Research shows most quality gains from RL post-training of diffusion transformers come from a constant shift: a fixed set of steering vectors (0.29M params, 65× fewer than RL-tuned LoRA) recovers 90.0–96.3% of reward gains across six RL-tuned image models and 89.8%/111.9% on two video models, and beats weight-based RL on OCR text rendering and PickScore on SD3.5-M. Policy-gradient RL largely steers the pretrained generator toward what it can already compute.
More from Multimodal
- Creator shows HeyGen-powered video edit: 'editing to another level' — aziz4ai · 2026-11-21
- AI-built oil painting engine: 2,300 commits in 8 days as Claude Code runs a 'studio' of sessions — FotoRe_store · 2026-11-21
- "AGI Before GTA6": AI-Generated Music Video Riffs on 13-Year GTA VI Wait — Yamapama · 2026-11-21
- Meta's AVTR adds visual imagination to multimodal reasoning, +4.5 pts across 16 benchmarks — liuzhuang1234 · 2026-11-21
- Genie 4 world model live on Comfy playground, early testers impressed — EnvironmentalDay7419 · 2026-11-21
- Roblox-style AI horror short built with CUT + local ComfyUI, workflow shared — rileygstaliger · 2026-11-21